CA2891714A1 - Bispecific antibodies - Google Patents

Abstract

The present invention relates to engineered heteromultimeric proteins, and more specifically, to methods for producing and purifying heterodimeric proteins, such as bispecific antibodies. Methods for producing and purifying such engineered heterodimeric proteins are also provided.

Description

BISPECIFIC ANTIBODIES
FIELD
The present invention relates to engineered bispecific antibodies and related polypeptides, multimeric forms thereof, and methods of making such proteins.
BACKGROUND
Antibodies having binding specificities for at least two different antigens, called bispecific antibodies (BsAbs), have been engineered. Unlike classical antibodies which comprise two identical heterodimer (i.e. a light chain portion and a heavy chain portion) "arms" wherein each arm comprises an antigen binding site (e.g. a Fab region), bispecific antibodies have different sequences in each of the two arms (e.g.
Fab regions) so that each arm of the Y-shaped molecule binds to a different antigen or different epitope of the same antigen.
By binding two different antigenic molecules or different epitopes of the same antigen, BsAbs may offer a wide variety of possible clinical applications as targeting agents for in vitro and in vivo diagnostics and immunotherapies. Bispecific antibodies may also be advantageous for in vitro or in vivo diagnoses of various disease states, including potentially cancer. For example, one arm of the BsAb may be engineered to bind a tumor-associated antigen and the other arm to bind a detectable marker.
BsAbs may also be used to direct a patient's cellular immune defense mechanisms to a tumor cell or an infectious agent (e.g. virally infected cells such as HIV or influenza virus; protozoa such as Toxoplasma gondii). For example, one may redirect immune modulated cytotoxicity by engineering one arm of the BsAb to bind to a desired target (e.g. a tumor cell or pathogen) and the other arm of the BsAb to bind to a cytotoxic trigger molecule, such as the T-cell receptor or a Fc gamma receptor, thereby potentially activating downstream immune effector pathways.
Using this strategy, BsAbs which bind to the Fc gamma RIII have been shown to mediate tumor cell killing by natural killer (NK) cell/large granular lymphocyte (LGL) cells in vitro and to prevent tumor growth in vivo. Alternatively, targeting two separate antigens or targets related to the therapeutic indication may enhance specificity and may reduce unwanted interaction, thereby potentially widening the therapeutic index.

=

- 2 -Although bispecific antibodies may posses certain advantages over canonical bivalent monospecific classical antibodies, use of bispecific antibodies has been hindered by the expense in obtaining BsAbs in sufficient quantity and purity.
To produce multispecific proteins, e.g. bispecific antibodies and other heterodimers or heteromultimers, it is desirable to use methods that favor formation of the desired heteromultimer over homomultimer(s). One method for obtaining Fc-containing BsAbs remains the hybrid hybridoma technique, in which two antibodies are co-expressed. However, this approach is inefficient with respect to yield and purity, the desired heteromultimer often being difficult to further purify from a relatively large level of contaminant comprising improperly paired polypeptide chains.
Other techniques to favor heteromultimer formation and reduce improper matching involve engineering sterically complementary mutations in multimerization domains at the CH3 domain interface, referred to as a "knobs-into-holes"
strategy as described by Ridgway et al. (US5731168) and Merchant et al. (US7183076).
Techniques that replace one or more residues that make up the CH3-CH3 interface in both CH3 domains with a charged amino acid for promoting the heterodimer formation have also been described by Strop et al.
(W02011/143545).
A recent review also discusses various approaches for overcoming chain association issues when generating bispecific antibodies (Klein et al., mAbs 4(6):
653-663 (2012)).
However, most of these techniques are directed to ensuring proper pairing of the heavy chain polypeptides and do not address the further matching of each light chain polypeptide with its corresponding heavy chain polypeptide to provide a functional antigen-binding site. Thus, production of desired bispecific antibodies remains a technically difficult and costly process not commercially feasible due to the high cost of goods.
Therefore, there is a need in the art for methods for engineering bispecific antibody fragments and/or full length BsAbs which enable the BsAbs to be expressed and recovered directly and/or efficiently from recombinant cell culture and/or which may be produced with efficient yields and purities at commercially reasonable costs.

- 3 -SUMMARY OF THE INVENTION
El. According to a first embodiment of the invention, there is provided a heterodimeric protein, comprising:
(i) a first CHCL domain (CHCL), comprising a first CH1 domain (CH1) and a first CL domain (CL), wherein the first CH1 and the first CL interact together at a first CHCL interface;
(ii) a second CHCL, comprising a second CH1 and a second CL, wherein the second CH1 and the second CL interacting together at a second CHCL
interface;
wherein the first CH1 differs from the second CH1 by at least one CH1 mutant residue in the first CH1; and the first CL differs from the second CL by at least one CL mutant residue in the first CL;
such that the CH1 mutant residue and the CL mutant residue of the first CHCL
interact with each other in preference to the corresponding residue positions on the second CHCL, the interacting mutant residues of the first CHI and first CL thereby forming a first complementary residue set.
Described below are a number of further embodiments (E) of this first embodiment of the invention, where for convenience El is identical thereto.
E2. The heterodimeric protein according to El ,wherein the second CH1 differs from the first CH1 by at least one CHI mutant residue in the second CHI; and the second CL differs from the first CL by at least one CL mutant residue in the second CL;
such that the CHI mutant residue and the CL mutant residue of the second CHCL
interact with each other in preference to the corresponding residue positions on the first CHCL, the interacting mutant residues of the second CH1 and second CL
thereby forming a second complementary residue set.
E3. A heterodimeric protein, comprising (i) a first CHI domain (CHI) and a first CL domain (CL), the first CH1 and the first CL interacting together at a first CHCL interface to form a first CHCL
domain (CHCL);
(ii) a second CH1 domain (CH1) and a second CL domain (CL), the second CH1

- 4 -and the second CL interacting together at a second CHCL interface to form a second CHCL domain (CHCL);
wherein the first CHI is engineered to differ from the second CH1 by at least one CH1 mutant residue in the first CH1; and the first CL is engineered to differ from the second CL by at least one CL
mutant residue in the first CL;
such that the CH1 mutant residue and the CL mutant residue of the first CHCL
interact with each other in preference to the corresponding residue positions on the second CHCL, the interacting mutant residues of the first CH1 and first CL thereby forming a first complementary residue set.
E4. The heterodimeric protein according to E3, wherein the second CH1 is engineered to differ from the first CH1 by at least one CH1 mutant residue in the second CH1; and the second CL is engineered to differ from the first CL by at least one CL mutant residue in the second CL; such that the CH1 mutant residue and the CL
mutant residue of the second CHCL preferentially interact with each other over the corresponding residue positions on the first CHCL, the interacting mutant residues of the second CH1 and second CL thereby forming a second complementary residue set.
E5. The heterodimeric protein according to any one of E1-E4, wherein the solvent accessible surface area of the first complementary residue set is less than 225A2 as measured using a 2.5A probe.
E6. The heterodimeric protein according to any one of E1-E5, wherein the solvent accessible surface area of the first complementary residue set is less than 220A2 as measured using a 2.5A probe.
E7. The heterodimeric protein according to any one of E1-E6, wherein the solvent accessible surface area of the first complementary residue set is less than 150A2 as measured using a 2.5A probe.
E8. The heterodimeric protein according to any one of E1-E7, wherein the solvent accessible surface area of the first complementary residue set is less than 100A2 as measured using a 2.5A probe.

- 5 -E9. The heterodimeric protein according to any one of El-E8, wherein the solvent accessible surface area of the first complementary residue set is less than 50A2 as measured using a 2.5A probe.
E10. The heterodimeric protein according to any one of El-E9, wherein the solvent accessible surface area of the first complementary residue set is less than 20A2 as measured using a 2.5A probe.
Eli. The heterodimeric protein according to any one of El-E10, wherein the solvent accessible surface area of the first complementary residue set is less than 10A2 as measured using a 2.5A probe.
E12. The heterodimeric protein according to any one of El-Ell, wherein the solvent accessible surface area of the first complementary residue set is less than 5A2 as measured using a 2.5A probe.
E12. The heterodimeric protein according to any one of El-E12, wherein the solvent accessible surface area of the first complementary residue set is less than 1A2 as measured using a 2.5A probe.
E13. The heterodimeric protein according to any one of El-E12, wherein the solvent accessible surface area of the second complementary residue set is less than as measured using a 2.5A probe.
E14. The heterodimeric protein according to any one of El-E13, wherein the solvent accessible surface area of the second complementary residue set is less than as measured using a 2.5A probe.
E15. The heterodimeric protein according to any one of El-E14, wherein the solvent accessible surface area of the second complementary residue set is less than as measured using a 2.5A probe.
E16. The heterodimeric protein according to any one of El-E15, wherein the solvent accessible surface area of the second complementary residue set is less than as measured using a 2.5A probe.
E17. The heterodimeric protein according to any one of El-E16, wherein the solvent accessible surface area of the second complementary residue set is less than as measured using a 2.5A probe.

- 6 -E18. The heterodimeric protein according to any one of E1-E17, wherein the solvent accessible surface area of the second complementary residue set is less than as measured using a 2.5A probe.
E19. The heterodimeric protein according to any one of E1-E18, wherein the solvent accessible surface area of the second complementary residue set is less than as measured using a 2.5A probe.
E20. The heterodimeric protein according to any one of E1-E19, wherein the solvent accessible surface area of the second complementary residue set is less than 5A2 as measured using a 2.5A probe.
E21. The heterodimeric protein according to any one of E1-E20, wherein the solvent accessible surface area of the second complementary residue set is less than 1A2 as measured using a 2.5A probe.
E21. The heterodimeric protein according to any one of E1-E20, wherein the solvent accessible surface area of the first or the second complementary residue set is about 0A2 as measured using a 2.5A probe.
E22. The heterodimeric protein according to any one of E1-E21, wherein the mutant residues of the first complementary residue set are different to the mutant residues of the second complementary residue set.
E23. The heterodimeric protein according to any one of E1-E22, wherein formation of the first CHCL and second CHCL preferentially occurs over formation of a CHCL
comprised of either the first CH1 and second CL, or second CH1 and first CL.
E24. The heterodimeric protein according to any one of E1-E23, wherein formation of the first CHCL and second CHCL preferentially occurs over formation of a CHCL
comprised of either the first CH1 and second CL, or second CH1 and first CL, by at least about 4-fold.
E25. The heterodimeric protein according to any one of E1-E24, wherein the first CH1 is attached to a first variable heavy domain (VH), and the first CL is attached to a first variable light domain (VL), and the second CH1 is attached to a second VH, and the second CL is attached to a second VL, E26. The heterodimeric protein according to to any one of E1-E25, wherein the -

- 7 -preferential formation of first CHCL and second CHCL does not rely on complementary pairing of the variable domains.
E27. The heterodimeric protein according to to any one of E1-E26, wherein formation of the first CHCL and second CHCL preferentially occurs over formation of a CHCL comprised of either the first CH1 and second CL, or second CH1 and first CL, by at least about 5-fold.
E28. The heterodimeric protein according to to any one of E1-E27, wherein formation of the first CHCL and second CHCL preferentially occurs over formation of a CHCL comprised of either the first CHI and second CL, or second CH1 and first CL, by -- at least about 6-fold.
E29. The heterodimeric protein according to to any one of E1-E28, wherein formation of the first CHCL and second CHCL preferentially occurs over formation of a CHCL comprised of either the first CH1 and second CL, or second CH1 and first CL, by at least about 8-fold.
-- E30. The heterodimeric protein according to to any one of E1-E29, wherein formation of the first CHCL and second CHCL preferentially occurs over formation of a CHCL comprised of either the first CH1 and second CL, or second CH1 and first CL, by at least about 10-fold.
E31. The heterodimeric protein according to to any one of E1-E30, wherein -- formation of the first CHCL and second CHCL preferentially occurs over formation of a CHCL comprised of either the first CH1 and second CL, or second CH1 and first CL, by at least about 15-fold.
E32. The heterodimeric protein according to to any one of E1-E31, wherein formation of the first CHCL and second CHCL preferentially occurs over formation of a -- CHCL comprised of either the first CH1 and second CL, or second CH1 and first CL, by at least about 20-fold.
E33. The heterodimeric protein according to to any one of E1-E32, wherein formation of the first CHCL and second CHCL preferentially occurs over formation of a CHCL comprised of either the first CH1 and second CL, or second CH1 and first CL, by -- at least about 25-fold.

'

- 8 -E34. The heterodimeric protein according to to any one of E1-E33, wherein formation of the first CHCL and second CHCL preferentially occurs over formation of a CHCL comprised of either the first CH1 and second CL, or second CH1 and first CL, by at least about 30-fold.
E35. The heterodimeric protein according to to any one of E1-E34, wherein formation of the first CHCL and second CHCL preferentially occurs over formation of a CHCL comprised of either the first CH1 and second CL, or second CH1 and first CL, by at least about 40-fold.
E36. The heterodimeric protein according to to any one of E1-E35, wherein formation of the first CHCL and second CHCL preferentially occurs over formation of a CHCL comprised of either the first CH1 and second CL, or second CH1 and first CL, by at least about 50-fold.
E37. The heterodimeric protein according to to any one of E1-E36, wherein formation of the first CHCL and second CHCL preferentially occurs over formation of a CHCL comprised of either the first CH1 and second CL, or second CH1 and first CL, by at least about 60-fold.
E38. The heterodimeric protein according to to any one of E1-E37, wherein formation of the first CHCL and second CHCL preferentially occurs over formation of a CHCL comprised of either the first CH1 and second CL, or second CH1 and first CL, by at least about 80-fold.
E39. The heterodimeric protein according to to any one of E1-E38, wherein formation of the first CHCL and second CHCL preferentially occurs over formation of a CHCL comprised of either the first CH1 and second CL, or second CHI and first CL, by at least about 90-fold.
E40. The heterodimeric protein according to to any one of E1-E39, wherein formation of the first CHCL and second CHCL preferentially occurs over formation of a CHCL comprised of either the first CH1 and second CL, or second CH1 and first CL, by at least about 100-fold.
E41. The heterodimeric protein according to any one of E25-E39, wherein the preferential formation of the first CHCL and second CHCL occurs in the absence of any =

- 9 -complementary pairing in the variable domains.
E42. The heterodimeric protein according to any one of E25-E41, wherein the first VH, first VL, first CH and first CL combined together form a first Fab, and the second VH, second VL, second CH1, and second CL combined together form a second Fab.
E43. The heterodimeric protein according to E42, wherein the preferential formation of first Fab and second Fab does not rely on complementary pairing of the variable domains.
E44. The heterodimeric protein according to any one of E42-E43, wherein the preferential formation of first Fab and second Fab occurs in the absence of any complementary pairing in the variable domains.
E44. The heterodimeric protein according to any one of E1-E43, wherein the preferential formation of first CHCL and second CHCL relies on complementary pairing of the complementary residue sets.
E45. The heterodimeric protein according to any one of E1-E46, wherein at least one of the CL domains is a kappa domain.
E46. The heterodimeric protein according to any one of E1-E45, wherein both the first CL and the second CL is a kappa domain.
E47. The heterodimeric protein according to any one of E1-E46, wherein the complementary residue sets comprise a positively or negatively charged residue in one domain, and either a polar residue, or an oppositely charged residue in the other domain.
E48. The heterodimeric protein according to any one of E1-E47, wherein the locations of the complementary residue sets are selected from the group consisting of:
CH1-124 and CL_176; (ii) CH1-188 and CL_178; (iii) CH1-143 and CL_178; (iv) and CL_131; (v) CH1-221 and CL_123; (vi) CH1-145 and CL_131; (vii) CH1-179 and CL_ 131; (viii) CH1-186 and CL_131; and (ix) CH1-188 and CL_133, according to Kabat numbering.
E49. The heterodimeric protein according to E48, wherein the mutation at the position is selected from the group consisting of W, H, K, R, Sand T, and the = CA 02891714 2015-05-15

- 10 -mutation at the CL position is selected from the group consisting of S, M, D
and E.
E50. The heterodimeric protein according to E49, wherein the mutation at the position is selected from the group consisting of E, and D, and the mutation at the CL
position is selected from the group consisting of H, K, and R.
E51. The heterodimeric protein according to any one of E49-E50, wherein the complementary residue sets further comprise one or more mutations selected from the group consisting of: CH1-143D, CH1-145S, CH1-186A, CH1-186E,CH1-188G, CH1-1435, CH1-190S, CH1-190I, CL-133S, CL-135I, CL-176G, CL-176M, and CL-178G, CL-1785.
E52. The heterodimeric protein according to any one of E1-E51, wherein the first and second complementary residue sets are selected from two of the following groups: CH1-124K, CL-176D, CH1-190S, CL-133S; (ii) CH1-124K, CL-176D, CL-133S;

(iii) CH1-124E, CL-176K; (iv) CH1-124E, CL-176K, CH1-188G; (v) CH1-188E, CL-178K, CH1-143E; (vi) CH1-188K, CL-178D, CH1-143D; (vii) CH1-143K, CL-178D; (viii) 143D, CL-178R; (ix) CH1-143K, CL-178D; (x) CH1-143D, CL-178K; (xi) CH1-143D, CL-178K, CL-176M; (xii) CH1-143E, CL-131R; (xiii) CH1-143R, CL-131E; (xiv) CH1-143R, CL-131E, CH1-186A; (xv) CH1-221D, CL-123K; (xvi) CH1-221D, CL-123K, CH1-190I, CL-135l; (xvii) CH1-145E, CL-131H; (xviii) CH1-143H, CH1-179D, CH1-186E, CL-131H;
(xix) CH1-145E, CL-131H; (xx) CH1-186E, CL-131H, CH1-145S; (xxi) CH1-143S, CL-131D, CH1-188W, CL-133S, CL-178S; (xxii) CH1-143S, CH1-188W, CL-133M, CL-176G, CL-178G; (xxiii) CH1-143H, CH1-179D, CH1-186E, CL-131H, CH-190I, CL-135I, (xxiv) CH-186E, CL-131H, CH-145S; (xxv) CH1-143S, CL-131D, CH1-188W, CL-133S, CL-176C; (xxvi) CH1-143S, CH1-188W, CL-133M, CL-178G, CL-176G; (xxvii) CH1-1435, CH1-188W, CL-131D.
E53. The heterodimeric protein according to any one of E1-E52, comprising an engineered disulfide bond between the first CH1 and the first CL, and or the second CH1 and the second CL.
E54. The heterodimeric protein according to E53, wherein the engineered disulfide bond is located at one or more of the following positions (i) CH1-122 and CL_123; (ii) CH1-139 and CL_116; and (iii) CH1-174 and CL_176.

-11 -E55. The heterodimeric protein according to any one of E53-E54, wherein a wild type disulfide bond has been removed, by mutating one or both of CH1-C230 and CL-214 to any residue except C, on the first CHCL and/or second CHCL.
E56. The heterodimeric protein according to E55, wherein the first and/or second CH1-C230 and first, and/or second CL-C214 are mutated to S.
E57. The heterodimeric protein according to any one of E1-E56, wherein the first CHCL comprises residues from one of the following groups: (i) CH1-124K, CL-176D, CH1-190S, CL-133S; (ii) CH1-124K, CL-176D, CL-133S; (iii) CH1-124E, CL-176K, CL-133S; (iv) CH1-124E, CL-176K, CH1-188G, CL-133S; (v) CH1-188E, CL-178K, CH1-143E; (v) CH1-188K, CL-178D, CH1-143D; (vi) CH1-143K, CL-178D; (vii) CH1-143D, CL-178R; (viii) CH1-143K, CL-178D; (ix) CH1-143D, CL-178K; (x) CH1-143D, CL-178K, CL-176M; (xi) CH1-143E, CL-131R; (xii) CH1-143R, CH 31E; (xiii) CH1-143R, CL-131E, CH1-186A; (xiv) CH1-221D, CL-123K; (xv) CH1-221D, CL-123K, CH1-190I, CL-1351, CH1-174C, CH1-230S, CL-176C, CL-214S; (xvi) CH1-145E, CL-131H; (xvii) 143H, CH1-179D, CH1-186E, CL-131H; (xviii) CH1-122C, CH1-145E, CH1-230S, CL-123C, CL-131H, CL-214S; (xix) CH1-186E, CL-131H, CH1-145S; (xx) CH1-143S, CL-131D, CH1-188W, CL-133S, CL-178S; (xxi) CH1-143S, CH1-188W, CL-133M, CL-176G, CL-178G; (xxii) CH1-143H, CH1-179D, CH1-186E, CL-131H, CH-190I, CL-135I, CH1-174C, CH1-230S, CL-176C, CL-214S; (xxiii) CH-186E, CL-131H, CH-145S, CH1-139C, CH1-230S, CL-116C, CL-214S; (xxiv) CH1-143S, CL-131D, CH1-188W, CL-1335, CL-178S, CH1-174C, CH1-230S, CL-176C, CL-214S; (xxv) CH1-143S, CH1-188W, CH1-122C, CH1-230S, CL-133M, CL-178G, CL-176G, CL-123C, CL-214S; (xxvi) CH1-143S, CH1-188W, CH1-122C, CH1-139C, CH1-174C, CH1-230S, CL-133S, CL-1785, CL-131D, CL-116C, CL-123C, CL-176C, CL-214S.
E58. The heterodimeric protein according to E57, wherein the second CHCL
comprises residues from one of groups i-xxvii, provided the first and second CHCL do not both comprises residues from the same group.
E59. The heterodimeric protein according to any one of E1-E58, wherein the first CH1 is connected to a first CH2 domain (CH2), which is connected to a first domain (CH3), and the second CH1 is connected to second CH2, which is connected

- 12 -to a second CH3.
E60. The heterodimeric protein according to E59, wherein the first CH3 and second CH3 comprises a first CH3 mutant residue and second CH3 mutant residue respectively, the first CH3 mutant residue and second CH3 mutant residues being engineered to differ from each other, and preferentially interact with each other and thereby form CH3 heterodimers over the formation of CH3 homodimers.
E61. The heterodimeric protein according to any one of E1-E60, wherein the first CH1 is attached to a first variable heavy domain (VH), and the first CL is attached to a first variable light domain (VL), and the second CH1 is attached to a second VH, and the second CL is attached to a second VL, and wherein the first VH comprises and VH-Q105.
E62. The heterodimeric protein according to any one of E1-E61, wherein the first CH1 is attached to a first variable heavy domain (VH), and the first CL is attached to a first variable light domain (VL), and the second CH1 is attached to a second VH, and the second CL is attached to a second VL, and wherein the second VH comprises VH-Q39 and VH-Q105.
E63. The heterodimeric protein according to any one of E1-E62, wherein the first CH1 is attached to a first variable heavy domain (VH), and the first CL is attached to a first variable light domain (VL), and the second CH1 is attached to a second VH, and the second CL is attached to a second VL, and wherein the first VL comprises:
(i) VL-Q38; and (ii) one of VL-Q1; VL-S1, VL-D1, VL-E1, VL-A1, or VL-N1; and (iii) one of VL-T42, VL-Q42, or VL-K42.
E64. The heterodimeric protein according to any one of E1-E63, wherein the first CH1 is attached to a first variable heavy domain (VH), and the first CL is attached to a first variable light domain (VL), and the second CH1 is attached to a second VH, and the second CL is attached to a second VL, and wherein the second VL comprises:
(i) VL-Q38; and (ii) one of VL-Q1; VL-S1, VL-D1, VL-E1, VL-A1, or VL-N1; and (iii) one of VL-T42, VL-Q42, or VL-K42.
E65. The heterodimeric protein according to any one of E1-E64, wherein the first CHCL comprises CH1-124K, CL-176D, CH1-190S, and CL-133S.

' CA 02891714 2015-05-15

- 13 -E66. The heterodimeric protein according to any one of E1-E65, wherein the second CHCL comprises CH1-124E, CL-176K, CH1-188G, and CL-133S.
E67. The heterodimeric protein according to any one of E1-E66, wherein the first CHCL comprises CH1-124K, CL-176D, CH1-190S, and CL-133S, and the second CHCL
comprises CH1-124E, CL-176K, CH1-188G, and CL-133S.
E68. A bispecific antibody comprising a heterodimeric protein as in any one of El-E67.
E69. The bispecific antibody as set forth in E66, wherein the first CHCL
comprises CH1-124K, CH 76D, CH1-190S, and CL-133S.
[70. The bispecific antibody as set forth in any one of E68-E69, wherein the second CHCL comprises CH1-124E, CL-176K, CH1-188G, and CL-133S.
E71. The bispecific antibody as set forth in any one of E68-E70, wherein the first CHCL comprises CH1-124K, CL-176D, CH1-190S, and CL-133S, and the second CHCL
comprises CH1-124E, CL-176K, CH1-188G, and CL-133S.
E72. A nucleic acid encoding the heterodimeric protein according to any one of El-E65, or a bispecific antibody according to any one of E68-E71.
E73. A vector comprising the nucleic acid according to E72.
E74. A cell comprising the nucleic acid according to E72, or comprising the vector according to E71.
E75. A method of making the heterodimeric protein, according to any one of El-E67, comprising: (i) cotransfecting a cell line with one or more vectors to express the first CH1, the first CL of the first CHCL; and the second CH1, and the second CL of the second CHCL; (ii) culturing the cell line under conditions to express the one or more vectors and that allow the first CHCL and second CHCL to assemble; and (iii) purifying the heterodimeric protein from the cell culture.
E76. The method of E75, wherein the cell line is cotransfected with vectors that express the first CH1, first CL, second CH1, and second CL in a 1:1:1:1 ratio.
BRIEF DESCRIPTION OF THE DRAWINGS
Figure 1 depicts potential products that may result from attempting to generate bispecific antibody via coexpression of 2 different antibody light chains and 2 different

- 14 -antibody heavy chains, where the CH3 interface is engineered using established technology to favor heterodirner formation, but the heavy/light chain interface is devoid of the mutations of the present invention. LC1 paired with HC1 provides a Fab arm binding one epitope, and LC2 paired with HC2 provides a Fab arm binding a distinct epitope, possibly on a different antigen. LC1 paired with HC2, or LC2 paired with HC1, results in a Fab with reduced or no binding to those epitopes. A:
correct pairing of a representative bispecific antibody, showing on the left arm, a combination of first VH and first VL (represented by heavy vertical stripes and light vertical dashes respectively) through preferential formation of first CHCL (facilitated by the interaction of a first complementary residue set, depicted by filled and unfilled circles), and on the right arm, a combination of second VH and second VL (represented by heavy diagonal hatching and light check pattern respectively) through preferential formation of the second CHCL (facilitated by the interaction of a second complementary residue set, depicted by filled and unfilled triangles). B: incorrect pairing of a bispecific antibody, showing on the left arm, a combination of first VH and second VL
(represented by heavy vertical stripes and light check pattern respectively) through formation of a third CHCL, and on the right arm, a combination of second VH
and first VL (represented by heavy diagonal hatching and light vertical dashes respectively) through formation of a fourth CHCL. C and D each show semi-functional bispecific antibodies, where either the left or right arm has paired correctly, and the other arm has paired incorrectly.
Figure 2 depicts the sequence of wild type human IgG1 CHI (panel A), Kappa CL (panel B), and Lambda CL domains. The amino acid residues are numbered according to the Kabat numbering scheme. A dash ("-") indicates an amino acid position that is occupied only in a different type of antibody domain or in a different species.
Figure 3 depicts non-reducing and reducing SDS-PAGE analysis of wild-type 29D7 monoclonal IgG1 antibody with native disulphide bridge (A), no disulphide bridge (B) or with disulphide bridges in novel positions (C-F) as detailed in Table 1 for clones Cys1, Cys3a, Cys3b and Cys6 respectively. M; molecular weight marker.

- 15 -Figure 4 depicts mass spectrometric analysis of constructs with engineered disulfides. Panel A: clone Cys_Beta, construct with wild type disulfides removed;
Panel B: clone Cys_1; Panel C: clone Cys_3a; Panel D: clone Cys_3b; Panel E:
clone Cys_6.
Figure 5 depicts x-ray crystal structures of the interface region involved in design Si. Within each panel, the CH1 domain is shown on top in dark gray with key residues rendered in ball and stick form. The CL domain is shown on the bottom in light gray with key resides rendered as tubes. Key interactions are indicated by dotted lines with distances in Angstroms. Panel A: the orientation of key residues and CL-S176, as well as the supporting residues CL-V133, CH1-S188, and CH1-in a native Fab arm comprising IgG1 CH1 and kappa CL. Panel B: Design to be used in one Fab arm of a standard two-arm antibody, with mutations CH1-L124K, CL-S176D, CL-V133S, and CH1-V190S. Panel C: Design to be used in the other Fab arm of a standard two-arm antibody, with mutations CH1-L124E, CL-5176K, CL-V1335, and CH1-5188G. Without wishing to be bound by any particular theory, when the mutations shown in B and C are introduced into each of the two Fab arms of an antibody, heavy/light mispairing may be disfavored by Lys/Lys or Asp/Glu charge repulsion, and/or correct pairing may be encouraged by Lys/Asp or Lys/Glu charge attraction. Panel A depicts PDB entry 3QQ9, while panels B and C are unpublished crystal structures.
Figure 6 depicts the results of mass spectrometric analysis of heterodimeric bispecific antibody Ab1/Ab2 comprising engineered favorable electrostatic interactions at the CH1/CL-Kappa interface of both Fab arms as described in Example 4 (panel A) and a control Ab1/Ab2 construct with native CH1/CL-Kappa interfaces (panel B). The novel electrostatic interaction mutations led to a significant reduction of incorrectly paired light chain in the isolated Fab fragments. Key: *
potential incomplete leader sequence processing; A Correctly paired (both H and L
chains) Ab2 Fab arm with a post-translational modification in the light chain.
Figure 7 depicts mass spectrometric analysis of bispecific antibody Abl/Ab2 Fc domain (panel A) and of a control Ab1/Ab2 containing HC heterodimerizing

- 16 -mutations but no Fab arm CH/CL interface mutations. Expected molecular weight of Fc (consisting of heavy chain from Ab 1 & 2) were detected in both cases whilst no heavy chain homodimers were detected.
Figure 8 depicts a graph showing the results from separation of bispecific antibody Ab1/Ab2 using anion-exchange chromatography. Anion-exchange chromatography was used to assess protein heterogeneity within the bispecific Ab1/Ab2 antibody preparation after protein A and preparative SEC
chromatography.
Analysis of the parental antibodies Ab1 and Ab2 are shown in panel A(i) and panel A(ii) respectively. Parental Ab1 displays an apparent single peak. Parental Ab2 shows a population of acidic and basic charge species which elute before and after the main peak, respectively. The heterodimeric bispecific Ab1/Ab2 antibody is shown in panel B. Fractions from Peak 1, Peak 2A and Peak 2B from the bispecific Ab1/Ab2 antibody (panel B) were analyzed by mass spectrometry.
Figure 9 depicts a graph showing a mass spectrometric analysis of Fab components from ion exchange fractionated heterodimeric bispecific antibody Ab1/Ab2 (derived in Figure 8B). Panel 9A shows that peak 2B from Figure 8B
contains enriched bispecific Ab1/Ab2 with correctly paired light chains in each Fab arm, but with a post translational modification in the Ab2 Fab arm. Panel 9B
shows Peak 2A from Figure 8B to has enriched incorrect light chain pairing (antibody heavy chain combined antibody 2 light chain), Panel 9C shows peak 1 from Figure 8B. This peak represents only correctly paired bispecific Ab1/Ab2 Fab arms with no post-translational modifications. Key: * potential incomplete leader sequence processing; A Ab2 Fab with a post-translational modification.
Figure 10 depicts a graph showing mass spectrometric analysis of dual arm Fab fragment of constructs C5XAb3-M1 & C5XAb3-M1-NEGATIVE. Significant reduction of incorrectly paired light chain between C5 & Ab3 was observed in construct C5XAb3-M1 compared to C5XAb3-M1-NEGATIVE. Key: * potential incomplete leader sequence processing.
Figure 11 depicts a graph showing mass spectrometric analysis of dual arm Fab fragment of constructs C5XAb3-M2 (panel 11B) and C5XAb3-M2-NEGATIVE

,

- 17 -(panel 11A). Significant reduction of incorrectly paired light chain between C5 & Ab3 was observed in construct C5XAb3-M2 compared to C5XAb3-M2-NEGATIVE. Key: *
potential incomplete leader sequence processing.
Figure 12 depicts a graph showing separation of bispecific antibodies using hydrophobic interaction chromatography. The parental antibodies Ab3 and C5 shown in panel 12A(1) and panel 12A(2), respectively, each display an apparent single peak. The heterodimerization approach M1 is shown in panel 12C and heterodimerization approach M2 is shown in panel 12B. The chromatograms on the left for both panel 12B(1) and 12C(1) show the incorporation of the heavy-chain heterodimerization mutations alone. The chromatograms on the right (panels 1213(2) and 12C(2)) show bispecific antibodies that contain both the heavy-chain and light-chain mutations described in Example 5. These results demonstrate a reduction in the heterogeneity of antibody produced with the incorporation of both CH1 and CL-Kappa mutations for correct light chain pairing and CH3 mutations compared with a bispecific (e.g., "NEGATIVE") comprising only the CH3/CH3 mutations.
Figure 13 depicts the interface region between a CH1 domain and a CL domain (from PDB entry 3QQ9). The view is along the interaction edge between the domains, with CH1 in dark gray on the left, and CL in light gray on the right.
Figure 14 depicts the interface region between a CH1 domain and a CL domain (from PDB entry 3QQ9) with a drawing style similar to Figure 13. This view highlights the regions of CH1 that interact with CL (backbone atoms of interacting residues shown with ball and stick rendering). The primary Ig-fold I3-strand regions are numbered 1 to 7 from N terminus to C terminus.
Figure 15 depicts the interface region between a CHI domain and a CL
domains (from PDB entry 3Q09) with a drawing style similar to Figure 13. This view highlights the regions of CL that interact with CH1 (backbone atoms of interacting residues shown with ball and stick rendering). The primary Ig-fold 13-strand regions are numbered 1 to 7 from N terminus to C terminus.
Figure 16 depicts a mostly buried, solvated pocket between the CH1 and CL
domains of PDB entry 3Q09. The light chain backbone, shown using a light gray

- 18 -ribbon, is in the front of the view, with the dark gray heavy chain backbone ribbon more to the back. Key water molecules defining this pocket are shown as spheres.
Figure 17 depicts graphs showing mass spectrometric analysis of dual arm Fab fragments designed to show the impact of subsets of the Si and Si_rev mutations. Panels A and B show original monospecific antibodies with no CH1/CL
mutations. A bispecific combining the two parent antibodies has significant mispairing in the absence of the Si and Si_rev mutations (Panel C) but nearly eliminated mispairing when Si and Si_rev are used (Panel D). Using various subsets of the Si and Si_rev mutations results in antibodies (Panels E and F) with reduced mispairing relative to Panel C, but still lower fidelity than the full Si and Si_rev design used in Panel D. The peaks corresponding to mispaired Fabs are labled as "Ab3H C5L"
and "C5H Ab3L" while the correct pairings are labeled as "C5 Fab" and "Ab3 Fab".
Figure 18 depicts a graph showing separation of bispecific antibodies using hydrophobic chromatography. In Panel A, the C5 and Ab3 antibodies are combined into a bispecific incorporating only CH3 mutations for heavy chain hetereodimerization, but no bispecific-favoring mutations in the heavy/light interface.
There are three major peaks, indicating a heterogenous sample. In Panel B, the Si and Si rev designs are added to the heavy/light interface to disfavor mispairing of the wrong heavy and light chains; the sample homogeneity is greatly improved. If some of the secondary supporting mutations of the Si and Si_rev designs are not utilized (Panels C and D), the sample has an intermediate level of heterogeneity.
Panels E
and F are controls showing the level of homogeneity observed with the monospecific versions of the two antibodies used to assemble the bispecific antibodies of Panels A-D.
Figure 19 depicts Differential Scanning Calorimetry (DSC) curves for bispecific Fabs with various combinations of mutations in each Fab arm, as described in Table 23. The solid thick lines indicate raw data, while the thin dotted lines indicate the results of fitting the raw data to a two transition or three transition model, as appropriate. As summarized in Table 24, all Fabs showed good stability with their lowest transition above 65 C.

- 19 -Figure 20 depicts mass spectrographic analysis of dual arm Fab fragments with various combinations of designs in each Fab arm, as enumerated in Table 23.
Bispecific antibodies with S1 in the Ab3 Fab arm and any of Ti, T2, T3, T4, or T9 in the C5 Fab arm displayed high fidelity of heavy/light chain pairing (Panels A-E). A
minor amount of mispairing 3%) was detected in the sample combining Si in one Fab arm with Si_rev in the other Fab arm (Panel F, mispair labeled as "C5 H
Ab3 L"). If one Fab arm (Panels G-H) or both Fab arms (Panel I) did not contain a bispecific-favoring design, larger amounts of mispaired Fab (19% or higher) were detected.
Figure 21 depicts separation of bispecific antibodies using hydrophobic interaction chromatography. The antibodies are enumerated in Table 23.
Bispecific antibodies with Si in the Ab3 Fab arm and any of Ti, T2, T3, T4, or T9 in the C5 Fab arm displayed high fidelity of heavy/light chain pairing (Panels A-E). A minor amount of mispairing is apparent as a small tail on the left side of the main peak.
This tail on the peak is slightly larger for Si on Ab3 paired with Si_rev on C5 (Panel F, see arrow). These results are consistent with the mass spectrographic analysis of Example 41 and Figure 20. If one Fab arm (Panels G-H) or both Fab arms (Panel I) did not contain a bispecific-favoring design, larger amounts of mispaired Fab were detected, as indicated by the presence of additional peaks. For reference, Panels J-K
show the corresponding profile of the monospecific Ab3 and C5 antibodies on which these bispecific designs were based.
Figure 22 depicts a bispecific antibody according to the invention. The domains are labelled as follows: 1-VL first variable light domain; 1-VH: first variable light domain. 1-CL: first constant light domain. 1-CH1: first constant heavy 1 domain.
1-CH2: first constant heavy 2 domain. 1-CH3: first constant heavy 3 domain. 2-VL
second variable light domain; 2-VH: second variable light domain. 2-CL: second constant light domain. 2-CH1: second constant heavy 1 domain. 2-CH2: second constant heavy 2 domain. 2-CH3: second constant heavy 3 domain. The first CHCL

and second CHCL domains are indicated between the braces (1-CHCL, and 2-CHCL
respectively) and encompass the respective CL and CH domains.The dotted oval lines

- 20 -capture the four domains (VL, VH, CL, CH1) that make up the first and second Fab (1-Fab and 2-Fab respectively). The first CHCL interface and second CHCL
interface are patterned in brickwork. Mutant residues in the CL and CH1 domains are represented by filled and unfilled circles and triangles (the set of filled and unfilled circles represent the complementary residue set of the first Fab and the set of filled and unfilled triangles represent the complementary residue set of the second Fab).
The 'knobs and holes' pairing of the first CH3 and second CH3 domains is represented by an arrow and ring.
This figure further illustrates the two semi-functional and one non-functional permutations that are sought to be avoided by the present invention. That is, the present inventions reduce the probability that a first CH1 (1-CH1) and a second CH (2-CL) will associate to form a third CHCL (Fig. 1B, left arm) compared with the favored pairings shown herein and in Figure 1A. Similarly, the present invention reduces the likelihood of formation of a fourth CHCL (comprising a second 2-CH1 and a 1-CL) as illustrated in Figure IC (right arm). Likewise, the present invention reduces the likelihood of formation of a non-functional antibody (e.g., Fig. 1B) comprising a third CHCL in one arm and a fourth CHCL in the other arm.
Figure 23: Chimeric TOA-1 antibody binds human TrkB
Figure 24: Chimeric TOA-1 antibody binds mouse TrkB
Figure 25: Humanized TOA-1 variants compete with biotinylated chimeric TOA-1 for binding to human TrkB
Figure 26: Humanized TOA-1 variants compete with biotinylated chimeric TOA-1 for binding to human TrkB
Figure 27: Humanized TOA-1 variants compete with biotinylated chimeric TOA-1 for binding to human TrkB
Figure 28: Humanized TOA-1 version 1.0/1.4 fully retains human TrkB binding properties relative to parental TOA-1 antibody Figure 29: Agonist activity of Anti-TrkB TOA-1 antibodies Figure 30: Humanized TOA-1 activates the TrkB signalling cascade Figure 31: The TOA-1 and BDNF binding sites on hTrkB overlap =

- 21 -Figure 32: TOA-1 binding to chimeric TrkB-TrkA receptors Figure 33:Anti-TrkB antibodies bind to mouse, cat and dog TrkB
Figure 34: TOA-1 antibodies do not bind to TrkA or TrkC
Figure 35: Humanized TOA-1 does not bind to p75 Figure 36: Humanized TOA-1 does not bind to p75 Figure 37: TOA-1 does not activate the TrkA or TrkC signaling cascades Figure 38. TAM-163 activates the Cre-luciferase reporter gene in hTrkB
cells Figure 39. TAM-163 does not activate the Cre-luciferase reporter gene in hTrkA-Cre and hTrkC-Cre cells Figure 40. TAM-163 activates hTrkB, but not hTrkA or hTrkC in the SHC1 recruitment assay Figure 41. Figure 1. TAM-163 activates TrkB-dependent phosphorylation events in hTrkB-Cre cells Figure 42 TAM-163 does not activate Trk-dependent phosphorylation events in hTrkA-Cre or hTrkC-Cre cells Figure 43. TAM-163 activates Irk-dependent phosphorylation events in human neuroblastoma SH-SY5Y cells Figure 44. TAM-163 induces internalization of TrkB in hTrkB-Cre and in human neuroblastoma SH-SY5Y cells Figure 45. TAM-163 induces degradation of TrkB in hTrkB-Cre and in human neuroblastoma SH-SY5Y cells Figure 46. TAM-163 does not bind to human p75NTR - FACS analysis Figure 47. TAM-163 does not bind human p75NTR - cell-based ELISA
Figure 48. TAM-163 binds to mouse, dog and cat TrkB with high affinity Figure 49. TAM-163 activates TrkB-dependent signaling in cells transfected with mouse TrkB
Figure 50. TAM-163 activates TrkB-dependent signaling in cells transfected with dog TrkB

- 22 -DETAILED DESCRIPTION
In some aspects, the invention relates to a heterodimeric protein comprising (i) a first CH1 domain (CH1) and a first CL domain (CL), the first CH1 and the first CL
interacting together at a first CHCL interface to form a first CHCL domain (CHCL); (ii) a second CH1 domain (CH1) and a second CL domain (CL), the second CH1 and the second CL interacting together at a second CHCL interface to form a second CHCL;
wherein the first CH1 is engineered to differ from the second CHI by at least one CH1 mutant residue in the first CH1; and the first CL is engineered to differ from the second CL by at least one CL mutant residue in the first CL; such that the CH1 mutant residue and the CL mutant residue of the first CHCL interact with each other in preference to the corresponding residue positions on the second CHCL, the interacting mutant residues of the first CH1 and first CL thereby forming a first complementary residue set.
In some aspects, the second CH1 is engineered to differ from the first CH1 by at least one CH1 mutant residue in the second CH1; and the second CL is engineered to differ from the first CL by at least one CL mutant residue in the second CL;
such that the CH1 mutant residue and the CL mutant residue of the second CHCL
preferentially interact with each other over the corresponding residue positions on the first CHCL, the interacting mutant residues of the second CH1 and second CL thereby forming a second complementary residue set.
The first CH1 may be engineered to differ from wild type CH1. The second CH1 may be engineered to differ from wild type CH1. The first CL may be engineered to differ from wild type CL. The second CL may be engineered to differ from wild type CL.
The first CH1 may comprise at least one CH1 mutant residue engineered to differ from the corresponding position on the second CH1. The first CL may comprise at least one CL mutant residue engineered to differ from the corresponding position on the second CL. The second CH1 may comprise at least one CH1 mutant residue engineered to differ from the corresponding position on the first CH1. The second CL
may comprise at least one CL mutant residue engineered to differ from the corresponding position on the first CL.

'

- 23 -In some aspects of the invention, the identity of the mutant residues of the first complementary residue set are different from the identity of the mutant residues of the second complementary residue set. In some aspects, the location of the mutant residues of the first complementary residue set are different from the location of the mutant residues of the second complementary residue set (locations according to Kabat numbering as described herein). In some aspects of the invention, the identity and location of the mutant residues of the first complementary residue set are different to the identity and location of the mutant residues of the second complementary residue set.
Preferential formation of heterodimers Providing a second complementary residue set in the second CHCL may further decrease the risk of mis-pairing of the domains. This strategy may be more effective when there is little overlap between the engineered mutations of the different domains. In some aspects, the first complementary residue set of the first CHCL are located at different positions relative to the location of the second complementary residue set of the second CHCL.
Accordingly, in some aspects of the invention, formation of the first CHCL and second CHCL preferentially occurs over formation of a CHCL comprised of either the first CH1 and second CL (hereinafter referred to as a third CHCL), or second CH1 and first CL (hereinafter referred to as a fourth CHCL).
Figure 1A and figure 22 illustrate a correctly paired antibody (comprising first CHCL
and second CHCL. Improperly paired domains are also depicted: a third CHCL
(comprising a first CHI and a second CL) is shown as the left arm of Figure 1B, and 1D, and a fourth CHCL (comprising a second CH1 and a first CL) is shown as the right arm of Figure 1B, and 1C. Similarity, switching the right hand and left hand light chains of Figure 22 would result in a non-functional antibody comprising third CHCL
and fourth CHCL.
In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL by at least about 4-fold. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the

- 24 -third and fourth CHCL by at least about 5-fold. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL
by at least about 6-fold. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL by at least about 7-fold. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL by at least about 8-fold. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL by at least about 9-fold. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL
by at least about 10-fold. In some aspects, formation of the first and second CHCL
may preferentially occur over formation of the third and fourth CHCL by at least about 12-fold. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL by at least about 15-fold.
In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL by at least about 20-fold. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL by at least about 25-fold. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL
by at least about 30-fold. In some aspects, formation of the first and second CHCL
may preferentially occur over formation of the third and fourth CHCL by at least about 35-fold. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL by at least about 40-fold.
In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL by at least about 50-fold. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL by at least about 60-fold. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL
by at least about 70-fold. In some aspects, formation of the first and second CHCL
may preferentially occur over formation of the third and fourth CHCL by at least about 75-fold. In some aspects, formation of the first and second CHCL may preferentially -

- 25 -occur over formation of the third and fourth CHCL by at least about 80-fold.
In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL by at least about 85-fold. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL by at least about 90-fold. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL
by at least about 95-fold. In some aspects, formation of the first and second CHCL
may preferentially occur over formation of the third and fourth CHCL by at least about 99-fold. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL by at least about 100-fold.
In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL by at least about 200-fold. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL by at least about 500-fold. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL by at least about 1000-fold.
In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 4 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 5 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 6 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 7 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 8 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 9 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 10 to about 1. In

- 26 -some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 12 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 15 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 20 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 25 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 30 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 35 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 40 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 45 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 50 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 55 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 60 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 65 to about 1.In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 70 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 75 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 80 to about 1. In

- 27 -some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 85 to about 1.1n some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 90 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 95 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 99 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 100 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 200 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 500 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 1000 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 2000 to about 1. In some aspects, formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 5000 to about 1.
The level of 'correct' heterodimer light chain pairing (i.e. first and second CHCL
formed) relative to 'incorrect' light chain pairing (i.e. third and fourth CHCL formed) may be measured by Liquid Chromatography Mass Spectrometry (LCMS). A
bispecific antibody preparation may be purified by protein A chromatography and preparative size exclusion chromatography to remove any aggregates or lower molecular weight components may be digested with LysC enzyme to release each Fab arm and the Fc as independent fragments (3 fragments total). LCMS may then be used to measure the empiral mass of each Fab arm and the Fc and values obtained are compared to the theoretical mass of the two possible correct Fab arms and the two possible incorrect Fab arms and for the Fc a comparison to theoretical =

- 28 -mass of homodimer vs heterodimer Fc is made. The signal intensity for each fragment can be converted to a % of total intensity of all fragments detected above background noise allowing for a ratio comparison of correct Fab product to incorrect Fab product. In a separate approach, post protein A bispecific antibody preparation elute can be fractionated using ion exchange or HIC chromatography and eluted fractions identified using LCMS. Identified peaks are then assigned % AUC from A280 measurements associated with the chromatography step. Ion exchange chromatography or hydrophobic interaction chromatography fractionate bispecific IgG
containing correct and incorrect light chain pairings based on differential charge or hydrophobicity properties. The % area under curve from the resulting A280 chromatograms can be used to quantitate the amount of correct product.
Solvent accessible surface area When introducing non-wild type human residues (such as the complementary residue sets herein; see below) into antibodies which may potentially be administered to a human, there is a risk that the human immune system will recognize the modified residues as foreign and generate antibodies against the administered agent (an anti-drug antibody or ADA response, which may result in faster clearance, reduced activity of circulating agent, or both). In order to be recognized by the ADA, the non-human residues of the administered antibody must be accessible to the ADA.
Minimizing the surface area accessible to the ADA may therefore be expected to reduce the ability of the ADA to interact with the administered antibody.
In some aspects, the solvent accessible surface area of the first complementary residue set is less than 50A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the second complementary residue set is less than 50A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the first complementary residue set is less than 225A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the second complementary residue set is less than 225A2 as measured using a 2.5A
probe. In some aspects, the solvent accessible surface area of the first complementary residue set is less than 220A2 as measured using a 2.5A probe.
In

- 29 -some aspects, the solvent accessible surface area of the second complementary residue set is less than 220A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the first complementary residue set is less than 150A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the second complementary residue set is less than 150A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the first complementary residue set is less than 120A2 as measured using a 2.5A
probe. In some aspects, the solvent accessible surface area of the second complementary residue set is less than 120A2 as measured using a 2.5A probe.
In some aspects, the solvent accessible surface area of the first complementary residue set is less than 100A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the second complementary residue set is less than as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the first complementary residue set is less than 80A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the second complementary residue set is less than 80A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the first complementary residue set is less than 50A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the second complementary residue set is less than as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the first complementary residue set is less than 40A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the second complementary residue set is less than 40A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the first complementary residue set is less than 30A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the second complementary residue set is less than as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the first complementary residue set is less than 20A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the second complementary residue set is less than 20A2 as measured using a 2.5A probe. In =

- 30 -some aspects, the solvent accessible surface area of the first complementary residue set is less than 10A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the second complementary residue set is less than as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the first complementary residue set is less than 5A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the second complementary residue set is less than 5A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the first complementary residue set is less than 2A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the second complementary residue set is less than 2A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of the first complementary residue set is less than 1A2 as measured using a 2.5A
probe. In some aspects, the solvent accessible surface area of the second complementary residue set is less than 1A2 as measured using a 2.5A probe.
In some aspects, the solvent accessible surface area is measured using the surface area algorithm in Maestro 9.6, 9.7, or 9.9 (Schrodinger, LLC.). The resolution may be 0.3. In some embodiments, the solvent accessible surface area of the first complementary residue set is less than 50A2 as measured using a 2.5A probe at high resolution (for example, a resolution of 0.3), using the surface area algorithm in Maestro 9.6, 9.7, or 9.9 (Schrodinger, LLC.).
It is well known in the art that mutation of a single side chain may improve antibody binding potency by an order of magnitude or more. For example, a His/Tyr substitution with an accessible surface area of - 90 A2 is known to cause a>
10-fold binding improvement of bevacizumab (J. Chem. Inf. Model. 53(11), 2937-50 (2013)).
However, it is well known in the art that even smaller surface changes may have similar effects. An alanine side chain has an accessible surface area of - 20 A2. A
mutation to alanine may be sufficient to change binding affinity between two proteins by greater than an order of magnitude. For example, see Mabs 3(5), 479-486 (2011).
Thus, a small mutated surface area may be sufficient to allow the immune system to produce an anti-drug antibody (ADA) which recognizes an engineered antibody, while ,

- 31 -potentially having significant selectivity against binding native human antibodies.
The solvent accessible surface area (SASA) is the surface of a biomolecule accessible to a solvent (typically water). SASA can be calculated by using the 'rolling ball' algorithm developed by Shrake & Rupley in 1973, which models a sphere approximating the size of the solvent molecule to 'probe' the surface of the molecule.
A typical value for the sphere radius is 1.4 A, as this corresponds to the approximate radius of a water molecule. However, a larger value (such as 2.5 A, as used herein) may be appropriate, when taking into account the experimental uncertainties in atom positions inherent in a crystal structure, or if the molecular entity, whose access to biomolecule's surface is in question, is larger than a water molecule (for example, the biomolecules of the potential host's immune system).
One aspect of the present invention is to provide a means of generating and maintaining bispecific heterogeneous antibodies or Fab fragments thereof through the use of engineered mutations in the CH1 and CL domains. However, introducing non-canonical residues into antibodies for potential in vivo use risks triggering a host immune response. It may therefore be advantageous to minimize the extent to which introduced or engineered residues to an antibody or Fab fragment thereof may potentially trigger a host immune response.
As noted above, in some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 50A2 as measured using a 2.5A probe.
In some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 45A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 40A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 35A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 30A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 25A2 as _

- 32 -measured using a 2.5A probe. In some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 20A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 15A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 10A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 9A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 8A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 7A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 6A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 5A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 4A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 3A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 2A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 1A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of a complementary residue set of the invention is less than 0.5A2 as measured using a 2.5A probe. In some aspects, the solvent accessible surface area of a complementary residue set of the invention is about 0A2 as measured using a 2.5A
probe.

- 33 -Variable domain In some aspects, the first CH1 is attached to a first variable heavy domain (VH), and the first CL is attached to a first variable light domain (VL), and the second CH1 is attached to a second VH, and the second CL is attached to a second VL. When combined, the first VH, first VL, first CH1, and first CL form a first Fab.
When combined, the second VH, second VL, second CH1, and second CL form a second Fab.
In some aspects, the first VH is connected to the first CH1, which in turn is connected to the first CH2, which in turn is connected to the first CH3, thereby forming a first heavy chain. In some aspects, the second VH is connected to the second CH1, which in turn is connected to the second CH2, which in turn is connected to the second CH3, thereby forming a second heavy chain.
In some aspects, the first VL is connected to the first CL, thereby forming a first light chain. In some aspects, the second VL is connected to the second CL, thereby forming a second light chain.
In some aspects, the invention provides for preferential formation of a first Fab and second Fab that does not rely on complementary pairing of the variable domains.
Where the preferential formation of heterodimeric protein domain interactions is discussed as not relying on complementary residue pairing of the variable domains, this means that the complementary pairing of, for example, a first CH1 and a first CL domain is sufficient to effect preferential formation of a first CHCL (or a first Fab). Additonal engineered residues in one or more of the variable or constant domains may provide additive effects to increase the fidelity of the preferential formation of the desired domain pairing.
In some aspects, the first complementary residue set may be necessary for preferential formation of the first CHCL. In some aspects, the first complementary residue set may be necessary for preferential formation of the first Fab. In some aspects, the second complementary residue set may be necessary for preferential formation of the second CHCL. In some aspects, the second complementary residue set may be necessary for preferential formation of the second Fab.
In some aspects, the first complementary residue set may be sufficent for . CA 02891714 2015-05-15 ,

- 34 -preferential formation of the first CHCL. In some aspects, the first complementary residue set may be sufficient for preferential formation of the first Fab. In some aspects, the second complementary residue set may be sufficent for preferential formation of the second CHCL. In some aspects, the second complementary residue set may be sufficient for preferential formation of the second Fab.
In some aspects, the invention provides for preferential formation of a first Fab and second Fab that does not rely on complementary pairing of the variable domains such that formation of the first and second CHCL may preferentially occur over formation of the third and fourth CHCL at a ratio of at least about 4 to about 1, and may occur at a ratio of at least a value selected from the group 4, 5, 6, 7, 8, 9, 10, 12, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 200, 500, 1000, 2000, and 5000 to 1.
There are some instances of known VLNH pairs that have a natural affinity with each other. Accordingly, in some aspects the invention provides for preferential formation of a first Fab and second Fab that does not rely on the any of the variable domains comprising engineered mutant residues that form complementary residue sets. In some aspects, the multimeric proteins of the invention do not comprise mutations in any of the variable domains that are engineered, which may increase complementary pairing above that of the non-engineered or wild type VLNH
framework sequences.
There may be multiple advantages realized by avoiding inserting mutant residues of complementary residue sets into variable domains. For example, it may be advantageous to use different germline frameworks for the variable region.
Sequence variations in each germ line present differing local environments for any mutations made in the variable domain; mutations which work in some frameworks may not work in other frameworks (for example, problems with expression, aggregation, stability, or other physical properties might occur). Also, mutations in the VLNH
interface (the area most likely to affect pairing specificity) are near CDRs and may affect the relative orientation of VL and VH in subtle ways that differ from antibody to antibody, and between frameworks. A subtle variation in VLNH orientation may be

- 35 -tolerated by some antibodies, but not others. In addition, mutating multiple regions of the protein surface (both variable and constant domains) may provide additional opportunities for the subject's immune system to recognize the antibody as foreign, and reject it via an anti-drug antibody response (ADA). Two possible results of an ADA response may be a faster rate of clearance of the administered agent from the subject, and neutralization of the agent's ability to bind its intended target (Jawa et.
al, Clin. Immunol. 149(3), 534-55 (2013)). In the development of bispecific antibodies, it may be desirable to take steps to minimize the probability of a potential subject's immune system mounting an ADA response. While there are some computational models for predicting T cell ADA response, accurate tools for conformational epitopes are lacking. Therefore, given the limited accuracy of in silico predictions, it may be preferable to limit modifications of high-fidelity bispecific IgG molecules to the CH1 and CL domains rather than mutating multiple domains as required by the Lewis et al.
method (see below).
In some aspects, the first VH comprises VH-Q39 (as in DP54 or DP75) or VH-Q105 (as in human J segments other than JH2). In some aspects, the second VH
comprises VH-Q39 (as in DP54 or DP75) orVH-Q105 (as in human J segments other than JH2).
In some aspects, the first VL comprises one or more of: (i) VL-Q38 (as in DPK9 or DPL16); and (ii) one of VL-Q1 (as in DPL7), VL-S1 (as in DPL16), VL-D1 (as in DPK9), VL-E1 (as in DPK23), VL-A1 (as in DPK3), or VL-N1 (as in DPK2); and (iii) one of VL-T42 (as in DPL7), VL-Q42 (as in DPL16), or VL-K42 (as in DPK9).
In some aspects, the second VL comprises one or more of: (i) VL-Q38 (as in DPK9 or DPL16) ; and (ii) one of VL-Q1 (as in DPL7), VL-S1 (as in DPL16), VL-D1 (as in DPK9), VL-E1 (as in DPK23), VL-A1 (as in DPK3), or VL-N1 (as in DPK2); and (iii) one of VL-T42 (as in DPL7), VL-Q42 (as in DPL16), or VL-K42 (as in DPK9).
In some aspects, the first VH comprises VH-Q39 (as in DP54 or DP75) and VH-Q105 (as in human J segments other than JH2). In some aspects, the second VH
comprises VH-Q39 (as in DP54 or DP75) and VH-Q105 (as in human J segments other than JH2). In some aspects, both the first and second VH comprise these

- 36 -residues.
In some aspects, the first VL comprises: (i) VL-038 (as in DPK9 or DPL16);
and (ii) one of VL-Q1 (as in DPL7), VL-S1 (as in DPL16), VL-D1 (as in DPK9), (as in DPK23), VL-A1 (as in DPK3), or VL-N1 (as in DPK2); and (iii) one of VL-T42 (as in DPL7), VL-Q42 (as in DPL16), or VL-K42 (as in DPK9).
In some aspects, the second VL comprises: (i) VL-Q38 (as in DPK9 or DPL16) ;
and (ii) one of VL-Q1 (as in DPL7), VL-S1 (as in DPL16), VL-D1 (as in DPK9), (as in DPK23), VL-A1 (as in DPK3), or VL-N1 (as in DPK2); and (iii) one of VL-T42 (as in DPL7), VL-Q42 (as in DPL16), or VL-K42 (as in DPK9).
In some aspects, both the first VL and second VL comprise the above residues.
Lewis et al. (Nat.Biotechno1.32, 191-98 (2014), or "Lewis publication"
hereafter) reported mutations in the CH1, CL, VL, and VH domains which attempted to address the issue of pairing light chains with the proper heavy chains. In a related patent application, W02014150973, bispecific antibodies are disclosed which all involve at least one mutation of a variable domain. The Lewis publication states: "Our method requires the introduction of multiple mutations into conserved framework regions of both variable and constant domains." The authors further noted that in their experience, "variable domains dominated the specific assembly of heavy chains and light chains". They hypothesized that during the protein folding pathway, the variable domains may "recognize one another first and drive the CL domain to interact with unfolded CH1", such that the heavy/light chain pairing is largely determined by interactions of VH and VL, before CHI and CL interact. That hypothesis would explain their observation that mutations in the variable region were required.
In contrast, the present invention provides heterodimeric proteins (e.g.
bispecific antibodies) which require no mutations of the CDRs or even the remainder of the variable region, and yet achieve high fidelity of chain pairing. Thus, relative to recent art in the field, specifically Lewis and W02014150973, the heterodimeric proteins and bispecific antibodies of the present invention are unexpected and may provide significant beneficial advantages.
As is known in the art, interactions between an antibody and its antigen are =

- 37 -driven primarily by the CDR loops. While not all CDR loops participate in antigen binding for all antigens, when designing a method of antibody engineering to try and achieve high fidelity bispecific chain pairing, mutation of positions within the CDRs and variable regionis a disadvantage due to the risk of negatively affecting antibody binding affinity. For cases involving the simultaneous production of multiple Fab sequences (or a bispecific IgG) rather than a single Fab, the various embodiments of W02014150973 all envision mutating the CDR2 region of the heavy chain as defined by Kabat ("the residue which is four amino acids upstream of the first residue of HFR3 according to Kabat" is mutated to glutamate, where HFR3 refers to framework 3 of the heavy chain). The heterodimeric proteins and bispecific antibodies of the present invention do not involve modification of the CDRs, and thus address this risk.
In addition, position 1 of the light chain variable region (which is mutated to Arg during production of four-chain mixtures according to the claims of W02014/150973A1) is near the CDR1 and CDR3 loops, which means that mutations at this position may also affect binding affinity to some antigens. In PDB
entry 4LLY, a crystal structure described in the Lewis publication, the side chain of position 1 is disordered beyond C13, but the backbone atoms are within 5 A of CDR L1 and within 6 A of CDR L3, and Cr3 is oriented towards the face of the Fab containing the majority of the CDR residues (ie, where antigen is expected to bind). In contrast, the heterodimeric proteins and bispecific antibodies of the present invention do not involve mutation of this position, or of any other position in the variable domain, thus addressing the risk of disturbing CDR positioning and/or antigen binding which exists when mutating nearby framework residues.
In principle, heterodimer-favoring mutations could be included in either of the major interface regions between the heavy and light chain, which are the interface between the CH1 and CL domains, and the interface between the variable heavy and variable light domains. However, as noted in part above, mutations in the interface are highly preferred for development of a robust bispecific platform.
Mutations in the variable domain interface may affect the conformation of the CDR
loops: because the CDR loops form part of the variable domain interface, they may

- 38 -interact (either directly or indirectly through nearby residues) with mutations made in the variable domains. If such interactions with heterodimer-enhancing mutations alter the CDR loop conformations in ways which affect antibody affinity, these heterodimer mutations may prove to be poor candidates for reliable use across a broad range of antibodies.
In addition, it is known that the relative orientation of the two variable domains is not constant among all antibodies; the angle between the two domains can vary by at least 30 degrees between antibodies (Abhinandan and Martin, Protein Eng Des Sel. 23(9), 689-97, (2010)). These changes necessarily alter the detailed pattern of contacts between residues in the variable domains, and correspondingly alter the range of amino acid substitutions that may be tolerated in the interface.
Given these facts, if variable domain mutations were used in a heterodimer-favoring platform design, it would be difficult to demonstrate robust and reliable applicability without testing a large number of examples covering the various CDR
conformations and variable domain orientation angles encountered in common practice in known antibody structures. Accordingly, generating heterodimers and bispecific anitbodies according to the present invention relies on modifying the CH1/CL interface. None of the embodiments of the present invention require pairing with modifications of the variable domains as an essential feature to achieve useful levels of pairing fidelity.
In one aspect, preferential formation of first Fab and second Fab relies on complementary pairing of the complementary residue sets.
In some aspects, preferential formation refers to the formation of a first Fab (or first CHCL) comprising the first CH1 and first CL to a greater extent than the formation of a Fab (or CHCL) comprising the first CH1 with a second CL, or a second CH1 with a first CL. In some aspects, preferential formation refers to the formation of a second Fab (or second CHCL) comprising the second CH1 and second CL to a greater extent than the formation of a Fab (or CHCL) comprising the first CH1 with a second CL, or a second CH1 with a first CL.
In some aspects, at least one of the CL domains is a kappa domain. In some

- 39 -aspects, at least one of the CL domains is a lambda domain. In some aspects, both of the CL domains are kappa domains. In some aspects, both of the CL domains are lambda domains. In some aspects, one of the CL domains is a kappa domain, and the other CL domain is a lambda domain.
In some aspects, the present invention provides for heterodimeric proteins and bispecific antibodies wherein formation of the first CHCL and second CHCL
preferentially occurs over formation of a CHCL comprised of either the first CH1 and second CL, or second CH1 and first CL, potentially by at least about an amount selected from the group consisting of 4-fold, 5-fold, 6-fold, 8-fold, 10-fold, 15-fold, 20-fold, 25-fold, 30-fold, 40-fold, 50-fold, 60-fold, 80-fold, 90-fold, 100-fold, 150-fold, and 200-fold.
The determination of correct CHCL pairing may be made by mass spectrometry analysis.
Complementary residue sets In some aspects, the complementary residue sets comprise a positively or negatively charged residue in one domain, and an oppositely charged residue in the other domain. In some aspects, the complementary residue sets comprise a positively charged residue in one domain, and negatively charged residue in the other domain. In some aspects, the complementary residue sets comprise a positively or negatively charged residue in one domain, and either a polar residue, or oppositely charged residue, in the other domain. Positively charged residues may be selected from the group consisting of H, K and R. Negatively charged residues may be selected form the group consisting of E and D. For the avoidance of doubt, negatively charged residues are said to be oppositely charged to positively charged residues, and vice versa. Polar residues may be selected from the group consisting of S, T, M, Q, N, W, and Y. Polar residues may be selected from the group consisting of S, T, M, Q, N, and W. Polar residues may be selected from the group consisting of S, T, M, Q, N, and Y. Polar residues may be selected from the group consisting of S, T, M, W, and Y. Polar residues may be selected from the group consisting of S, T, M, W, and Y. Polar residues may be selected from the group consisting of S, T, M, and W.
Polar = CA 02891714 2015-05-15

- 40 -residues may be selected from the group consisting of S, M, W, and Y. Polar residues may be selected from the group consisting of S, M, and W. Polar residues may be selected from the group consisting of S and T. In some aspects, M is not considered to be a polar residue.
For example, the CH1 mutant residue may comprise a positively or negatively charged residue, and the CL mutant residue may comprise either a polar residue, or an oppositely charged residue. The CL mutant residue may comprise a positively or negatively charged residue, and the CH1 mutant residue may comprise either a polar residue, or an oppositely charged residue. The CL mutant residue may comprise a -- positively charged residue, and the CH1 mutant residue may comprise a negatively charged residue. The CH1 mutant residue may comprise a positively charged residue, and the CL mutant residue may comprise a negatively charged residue.
In some aspects of the invention, the complementary residue sets may comprise a CH1 mutant residue and a CL mutant residue whose oppositely charged side chains -- promote electrostatic interaction. Favorably, the altered charge polarity of the respective CH1 and CL domains resulting from the engineered mutant residues supports the formation of the first or second Fab, and similarly, a repulsive charge interaction resulting from one or more of the engineered mutant residues suppresses the formation of the third or fourth Fab.
In some aspects, the locations of the complementary residue sets are selected from the group consisting of: (i) CH1-124 and CL_176; (ii) CH1-188 and CL_178;
(iii) CH1-143 and CL_178; (iv) CH1-143 and CL_131; (v) CH1-221 and CL_123; (vi) CH1-and CL..131; (vi) CH1-179 and CL_131; (vii) CH1-186 and CL_131; and (viii) CH1-143 and CL_133, according to Kabat numbering as defined herein.
In some aspects, the complementary residue set comprises CH1-124 and CL_176.
In some aspects, the complementary residue set comprises CH1-188 and CL_178.
In some aspects, the complementary residue set comprises CH1-143 and C1..178. In some aspects, the complementary residue set comprises CH1-143 and CL_131. In some aspects, the complementary residue set comprises CH1-221 and CL_123. In -- some aspects, the complementary residue set comprises CH1-145 and CL_131.
In . CA 02891714 2015-05-15 ,

-41 -some aspects, the complementary residue set comprises CH1-179 and CL_131. In some aspects, the complementary residue set comprises CH1-145, CH1-179, CH1-186 and CL_131. In some aspects, the complementary residue set comprises CH1-143, CH1-179, CH1-186, and CL_131. In some aspects, the complementary residue set comprises CH1-186 and CL-131. In some aspects, the complementary residue set comprises CH1-143 and CLA 33.
In some aspects, the mutation at the CH1 position is selected from the group consisting of W, H, K, R, S and T, and the mutation at the CL position is selected from the group consisting of S, M, D and E.
In some aspects, the mutation at the CH1 position is selected from the group consisting of E, and D, and the mutation at the CL position is selected from the group consisting of H, K, and R.
In some aspects, one or more of the complementary residue sets further comprise one or more further mutations.
In some aspects, one or more of the complementary residue sets comprise one or more further mutations selected from the group consisting of: CH1-143D, CH1-145S, CH1-186A, CH1-186E,CH1-188G, CH1-188W, CH1-190S, CH1-190I, CL-133S, CL-1351, CL-176G, CL-176M, and CL-178S.
In some aspects, one or more of the complementary residue sets comprise further mutations located at one or more positions selected from the group consisting of:
CH1-143, CH1-145, CH1-186, CH1-188, CH1-188, CH1-190, CH1-190, CL-133, CL-135, CL-176, CL-176, and CL-178, according to Kabat numbering as described herein.
In some aspects, one or more of the complementary residue sets comprise a further CH1 mutant residue at CH1-143. The mutant residue at CH1-143 may be selected from the group consisting of H, K, R, E, and D. The mutant residue at 143 may be selected from the group consisting of E, and D. The mutant residue at CH1-143 may be E. The mutant residue at CH1-143 may be D.
In some aspects, one or more of the complementary residue sets comprise a further CH1 mutant residue at CH1-145. The mutant residue at CH1-145 may be selected from the group consisting of S, T, M, Q, N, E, D, W, or Y. The mutant =

- 42 -residue at CH1-145 may be selected from the group consisting of S, T, M, Q, N, E, or D. The mutant residue at CH1-145 may be selected from the group consisting of S, T, M, Q, or N. The mutant residue at CH1-145 may be selected from the group consisting of S, T, or M. The mutant residue at CH1-145 may be S. The mutant residue at CH1-145 may be T.
In some aspects, one or more of the complementary residue sets comprise a further CH1 mutant residue at CH1-186. The mutant residue at CH1-186 may be selected from the group consisting of G, A, L, V, I, W, F, or Y. The mutant residue at CH1-186 may be selected from the group consisting of G, A, L, V, I, or W. The mutant residue at CH1-186 may be selected from the group consisting of G, A, L, V, oil. The mutant residue at CH1-186 may be selected from the group consisting of G, A, V, or L. The mutant residue at CH1-186 may be selected from the group consisting of G, A, or V. The mutant residue at CH1-186 may be selected from the group consisting of G, or A. The mutant residue at CH1-186 may be selected from the group consisting of A, or W. The mutant residue at CH1-186 may be selected from the group consisting of F, Y, or W. The mutant residue at CH1-186 may W. The mutant residue at CH1-186 may A.
In some aspects, one or more of the complementary residue sets comprise a further CH1 mutant residue at CH1-188. The mutant residue at CH1-188 may be selected from the group consisting of G, A, L, V, I, W, F, or Y. The mutant residue at CH1-188 may be selected from the group consisting of G, A, L, V, I, or W. The mutant residue at CH1-188 may be selected from the group consisting of G, A, L, V, or I. The mutant residue at CH1-188 may be selected from the group consisting of G, A, V, or L. The mutant residue at CH1-188 may be selected from the group consisting of G, or A. The mutant residue at CH1-188 may be selected from the group consisting of G, A, or W. The mutant residue at CH1-188 may be selected from the group consisting of G, or W. The mutant residue at CH1-188 may be selected from the group consisting of F, Y, or W. The mutant residue at CH1-188 may be W. The mutant residue at 188 may be A. The mutant residue at CH1-188 may be G.
In some aspects, one or more of the complementary residue sets comprise a =

- 43 -further CH1 mutant residue at CH1-190. The mutant residue at CH1-190 may be selected from the group consisting of S, T, I, L. The mutant residue at CH1-190 may be selected from the group consisting of I or L. The mutant residue at CH1-190 may be selected from the group consisting of S or T. The mutant residue at CH1-190 may be selected from the group consisting of S or I. The mutant residue at CHI-190 may be T. The mutant residue at CH1-190 may be L. The mutant residue at CH1-190 may be I. The mutant residue at CH1-190 may be S.
In some aspects, one or more of the complementary residue sets comprise a further CL mutant residue at CL-133. The mutant residue at CL-133 may be selected from the group consisting of S, T, Q or M. The mutant residue at CL-133 may be S.
The mutant residue at CL-133 may be T. The mutant residue at CL-133 may be M.
The mutant residue at CL-133 may be Q.
In some aspects, one or more of the complementary residue sets comprise a further CL mutant residue at CL-135. The mutant residue at CL-135 may be selected from the group consisting of I, T, or M. The mutant residue at CH 35 may be I.
In some aspects, one or more of the complementary residue sets comprise a further CL mutant residue at CL-176. The mutant residue at CL-135 may be selected from the group consisting of G, A, V, I, L, M, N. or T. The mutant residue at may be selected from the group consisting of G, A, V, I, L, or M. The mutant residue at CL-176 may be selected from the group consisting of G, A, V, L, or M. The mutant residue at CL-176 may be selected from the group consisting of G, A, V, or M.
The mutant residue at CL-176 may be selected from the group consisting of G, A, or M.
The mutant residue at CL-176 may be selected from the group consisting of G, or M.
The mutant residue at CL-176 may be G. The mutant residue at CL-176 may be A.
The mutant residue at CL-176 may be M. The mutant residue at CL-176 may be N.
In some aspects, one or more of the complementary residue sets comprise a further CL mutant residue at CL-178. The mutant residue at CL-135 may be selected from the group consisting of G, S, V, or A. The mutant residue at CL-135 may be S.
In some aspects, wherein the first and second complementary residue sets are selected from two of the following groups: (i) CH1-124K, CL-176D; (ii) CH1-124K, CL-=
,

- 44 -176D, CH1-190S, CL-133S; (iii) CH1-124K, CL-176D, CL-133S; (iv) CH1-124E, CL-176K; (v) CH1-124E, CL-176K, CH1-188G; (vi) CH1-188E, CL-178K, CH1-143E; (vii) CH1-188K, CL-178D, CH1-143D; (viii) CH1-143K, CL-178D; (ix) CH1-143D, CL-178R;

(x) CH1-143K, CL-178D; (xi) CH1-143D, CL-178K; (xii) CH1-143D, CL-178K, CL-176M;
(Xiii) CH1-143E, CL-131R; (xiv) CH1-143R, CL-131E; (xv) CH1-143R, CL-131E, CH1-186A; (xvi) CH1-221D, CL-123K; (xvii) CH1-221D, CL-123K, CH1-190I, CL-1351;
(xviii) CH1-145E, CL-131H; (xvix) CH1-143H, CH1-179D, CH1-186E, CL-131H; (xix) CH1-145E, CL-131H; (xx) CH1-186E, CL-131H, CH1-145S; (xxi) CH1-143S, CL-131D, CH1-188W, CL-133S, CL-178S; (xxii) CH1-143S, CH1-188W, CL-133M, CL-176G, CL-178G;
(xxiii) CH1-143H, CH1-179D, CH1-186E, CL-131H, CH-190I, CL-1351, (xxiv) CH-186E, CL-131H, CH-145S; (xxv) CH1-143S, CL-131D, CH1-188W, CL-133S, CL-176C; (xxvi) CH1-143S, CH1-188W, CL-133M, CL-178G, CL-176G; (xxvii) CH1-143S, CH1-188W, CL-131D.
Novel disulfide linkage In some aspects, the invention provides for a novel disulfide bond between the first CH1 and the first CL, and/or the second CH1 and the second CL. The novel disulfide bond may be located at one or more of the following positions (i) and CL_123; (ii) CH1-139 and CL_116; and (iii) CH1-174 and CL_176.
The wild type disulfide bond may be removed, by mutating one or both of CH1-C230 and CL-214 to any residue except C, on either or both of the first CHCL
and/or second CHCL. In some aspects, the CL-C214 is deleted in either or both of the first and/or second CHCL. In some aspects, the CH1-C230 is deleted in either or both of the first and/or second CHCL.
In some aspects, the first and/or second CH1-C230 and first, and/or second CL-C214 are mutated to S. In some aspects, the first CHCL comprises CH1-C230S
and CL-C214S, and further comprises one or more of the following residue pairs:
CH1-122C and CL_123C; CH1-139C and CL_116C; and CH1-174C and CL_176C. In some aspects, the second CHCL comprises CH1-C230S and CL-C214S, and further comprises one or more of the following residue pairs: CH1-122C and CL_123C;

139C and CL_116C; and CH1-174C and CL_176C. Favorably, the first CHCL and

- 45 -second CHCL do not comprise novel cytokine mutations located at the same corresponding positions.
In some aspects, wherein a given CHCL comprises CH-174C and CL-176C, the given CHCL further comprises CH-190I and CL-135I.
CH/CL mutations In some aspects, the first CHCL and/or second CHCL comprises residues from one of the following groups: (i) CH1-124K, CL-176D, CH1-190S, CL-133S; (ii) 124E, CL-176K, CH1-188G, CL-133S; (iii) CH1-124K, CL-176D, CL-133S; (iv) CH1-124E, CL-176K, CL-133S; (v) CH1-188E, CL-178K, CH1-143E; (vi) CH1-188K, CL-178D, CH1-143D; (vii) CH1-143K, CL-178D; (viii) CH1-143D, CL-178R; (ix) CH1-143K, CL-178D; (x) CH1-143D, CL-178K; (xi) CH1-143D, CL-178K, CL-176M; (xii) CH1-143E, CL-131R; (xiii) CH1-143R, CL-131E; (xiv) CH1-143R, CL-131E, CH1-186A; (xv) CH1-221D, CL-123K; (xvi) CH1-221K, CL-123K, CH1-190I, CL-135I, CH1-174C, CH1-230S, CL-176C, CL-214S; (xvii) CH1-145E, CL-131H; (xviii) CH1-143H, CH1-179D, CH1-186E, CL-131H; (xix) CH1-122C, CH1-145E, CH1-230S, CL-123C, CL-131H, CL-214S; (xx) CH1-186E, CL-131H, CH1-145S; (xxi) CH1-143S, CL-131D, CH1-188W, CL-133S, CL-178S; (xxii) CH1-143S, CH1-188W, CL-133M, CL-176G, CL-178G; (xxiii) CH1-143H, CH1-179D, CH1-186E, CL-131H, CH-190I, CL-135I, CH1-174C, CH1-230S, CL-176C, CL-214S; (xxiv) CH-186E, CL-131H, CH-145S, CH1-139C, CH1-230S, CL-116C, CL-214S; (xxv) CH1-143S, CL-131D, CH1-188W, CL-133S, CL-178S, CH1-174C, CH1-230S, CL-176C, CL-214S; ()o(vi) CH1-221D, CL-123K, CH1-190I, CL-135I, CH1-174C, CHI-230S, CL-176C, CL-214S; (xxvii) CHI-143S, CHI-188W, CH1-122C, CH1-139C, CHI-174C, CHI-230S, CL-133S, CL-178S, CL-131D, CL-116C, CL-123C, CL-176C, CL-214S.
In some embodiments, the first and second Fab do not both comprise residues from the same group.
In some aspects, the first CHCL and/or second CHCL comprises the residues CH1-124K, CL-176D, CH1-190S, and CL-133S. In some aspects, the first CHCL
and/or second CHCL comprise the residues CH1-124K, and CL-176D. In some aspects, the first CHCL and/or second CHCL comprises the residues CH1-124K, CL-176E. In some

-46 -aspects, the first CHCL and/or second CHCL comprise the residues CH1-124R, CL-176D. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-124R, CL-176E.
In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-124E, CL-176K, CH1-188G, and CL-133S. In some aspects, the first CHCL
and/or second CHCL comprise the residues CH1-124E, and CL-176K. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-124E, and CL-176R. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-124D, and CL-176K. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-124D, and CL-176R.
In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-188E, CL-178K, and CH1-143E. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-188E, and CL-178K. In some aspects, the first CHCL
and/or second CHCL comprise the residues CH1-188D, and CL-178K. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-188E, CL-178R. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-188D, CL-178R.
In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-188K, CL-178D, and CH1-143D. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-188K, CL-178D. In some aspects, the first CHCL
and/or second CHCL comprise the residues CH1-188R, CL-178D. In some aspects, the first CHCL and/or second CHCL comprise the residues CHI-188K, CL-178E. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-188R, CL-178E.
In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143K, and CL-178D. In some aspects, the first CHCL and/or second CHCL
comprise the residues CH1-143K, and CL-178E. In some aspects, the first CHCL
and/or second CHCL comprise the residues CH1-143R, and CL-178D. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143R, and CL-178E.

- 47 -In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143D, and CL-178R. In some aspects, the first CHCL and/or second CHCL
comprise the residues CH1-143E, and CL-178R. In some aspects, the first CHCL
and/or second CHCL comprise the residues CH1-143D, and CL-178K. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143E, and CL-178K.
In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143D, CL-178K, and CL-176M.
In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143E, and CL-131R. In some aspects, the first CHCL and/or second CHCL
comprise the residues CH1-143D, and CL-131R. In some aspects, the first CHCL
and/or second CHCL comprise the residues CH1-143E, and CL-131K. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143D, and CL-131K.
In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143R, CL-131E, and CH1-186A. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143R, and CL-131E. In some aspects, the first CHCL
and/or second CHCL comprise the residues CH1-143K, and CL-131E. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143R, and CL-131D. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143K, and CL-131D.
In some aspects, the first CHCL and/or second CHCL comprise the residues CHI-221D, CL-123K, CH1-190I, CL-1351, CH1-174C, CH1-230S, CL-176C, and CL-214S. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-221D, and CL-123K. In some aspects, the first CHCL and/or second CHCL
comprise the residues CH1-221E, and CL-123K. In some aspects, the first CHCL
and/or second CHCL comprise the residues CHI-221D, and CL-123R. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-221E, and CL-123R.
In some aspects, the first CHCL and/or second CHCL comprise the residues . CA 02891714 2015-05-15 ,

-48 -CH1-145E, and CL-131H. In some aspects, the first CHCL and/or second CHCL
comprise the residues CHI-145D, and CL-131H. In some aspects, the first CHCL
and/or second CHCL comprise the residues CH1-145E, and CL-131K. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-145E, and CL-131R. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-145D, and CL-131K. In some aspects, the first CHCL and/or second CHCL
comprise the residues CH1-145E, and CL-131H.
In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143H, CH1-179D, CH1-186E, and CL-131H. In some aspects, the first CHCL
and/or second CHCL comprise the residues CH1-143E, CH1-179D, CH1-186E, and CL-131H. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143D, CH1-179D, CH1-186D, and CL-131H. In some aspects, the first CHCL
and/or second CHCL comprise the residues CH1-143H, CH1-179D, CH1-186D, and CL-131H. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143H, CH1-179E, CH1-186D, and CL-131H. In some aspects, the first CHCL
and/or second CHCL comprise the residues CH1-143H, CH1-179E, CH1-186E, and CL-131H. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143H, CH1-179E, CH1-186D, and CL-131H. In some aspects, the first CHCL
and/or second CHCL comprise the residues CH1-143H, CH1-179E, CH1-186E, and CL-131H.
In some aspects, the first CHCL and/or second CHCL comprise the residues CHI-145E, CL-131H, CHI-122C, CHI-230S, CL-123C, and CL-214S.
In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-186E, CL-131H, and CHI-145S. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-186E, and CL-131H. In some aspects, the first CHCL
and/or second CHCL comprise the residues CH1-186D, and CL-131H.
In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143S, CL-131D, CH1-188W, CL-133S, and CL-178S. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143S, CH1-188W, and CL-131D. In some aspects, the first CHCL and/or second CHCL comprise the residues =

- 49 -CH1-143T, CH1-188W, and CL-131D. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143S, CHI-188W, and CL-131E. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143T, CH1-188W, and CL-131E.
In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143S, CH1-188W, CL-133M, CL-176G, and CL-178G. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143S, CH1-188W, and CL-133M. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143T, CH1-188W, and CL-133M.
In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143H, CH1-179D, CH1-186E, CL-131H, CH-190I, CL-135I, CH1-174C, CH1-230S, CL-176C, and CL-214S. In some aspects, the first CHCL and/or second CHCL
comprise the residues CH1-143H, CH1-179D, CH1-186E, and CL-131H. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143H, CH1-179E, CH1-186E, and CL-131H. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143H, CH1-179D, CH1-186D, and CL-131H. In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143H, CHI-179E, CH1-186D, and CL-131H.
In some aspects, the first CHCL and/or second CHCL comprise the residues CH-186E, CL-131H, CH-145S, CH1-139C, CH1-230S, CL-116C, and CL-214S.
In some aspects, the first CHCL and/or second CHCL comprise the residues CHI-143S, CL-131D, CHI-188W, CL-133S, CL-178S, CHI-174C, CHI-230S, CL-176C, and CL-214S.
In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143S, CH1-1 88W, CL-133M, CL-178G, CL-176G, CH1-122C, CH1-230S, CL-123C, and CL-214S.
In some aspects, the first CHCL and/or second CHCL comprise the residues CH1-143S, CL-131D, CHI-188W, CL-133S, CL-178S, CH1-122C, CH-139C, CH-174C, CH1-230S, CL-116C, CL-123C, CL-176C, and CL-214S.
In some aspects, the invention comprises a CH1 domain comprising a

- 50 -sequence identical to SEQ ID NO:1, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH1 domain comprising a sequence identical to SEQ ID NO:2, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH1 domain comprising a sequence identical to SEQ ID NO:3, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH1 domain comprising a sequence identical to SEQ ID NO:4, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH1 domain comprising a sequence identical to SEQ ID
NO:5, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH1 domain comprising a sequence identical to SEQ ID

NO:6, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH1 domain comprising a sequence identical to SEQ ID NO:7, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CHI domain comprising a sequence identical to SEQ ID NO:8, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%.
In some aspects, the invention comprises a CH1 domain comprising a sequence identical to SEQ ID NO:33, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH1 domain comprising a sequence identical to SEQ ID NO:34, by at least an amount selected

- 51 -from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH1 domain comprising a sequence identical to SEQ ID N0:35, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH1 domain comprising a sequence identical to SEQ ID N0:36, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH1 domain comprising a sequence identical to SEQ ID
N0:37, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH1 domain comprising a sequence identical to SEQ ID NO:38, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH1 domain comprising a sequence identical to SEQ ID N0:39, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH1 domain comprising a sequence identical to SEQ ID N0:40, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH1 domain comprising a sequence identical to SEQ ID N0:41, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%.
In some aspects, the invention comprises a CL domain comprising a sequence identical to SEQ ID N0:9, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CL domain comprising a sequence identical to SEQ ID N0:10, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and

- 52 -99%. In some aspects, the invention comprises a CL domain comprising a sequence identical to SEQ ID NO:11, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CL domain comprising a sequence identical to SEQ ID NO:12, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%.
In some aspects, the invention comprises a CL domain comprising a sequence identical to SEQ ID NO:24, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CL domain comprising a sequence identical to SEQ ID NO:25, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CL domain comprising a sequence identical to SEQ ID NO:26, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CL domain comprising a sequence identical to SEQ ID NO:27, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CL domain comprising a sequence identical to SEQ ID NO:28, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CL domain comprising a sequence identical to SEQ ID NO:29, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CL domain comprising a sequence identical to SEQ ID NO:30, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%.In some aspects, the invention comprises a CL domain comprising a sequence identical to SEQ ID NO:31, by at least an amount selected from the group consisting

- 53 -of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CL domain comprising a sequence identical to SEQ ID NO:32, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%.
Modification to constant domains of antibodies to produce heterodimers are disclosed in US5731168, W02009089004, and W02011143545, each of whose contents is herein incorporated in its entirety.
CH2 and CH3 domains In some aspects, the first CH1 is connected to a first CH2 domain (CH2), and the second CH1 is connected to second CH2. The first and second CH2 may each comprise a first and second CH2 mutant residue respectively, the first and second CH2 mutant residues being engineered to differ from each other, and preferentially interact with each other and thereby form CH2 heterodimers preferentially over the formation of CH2 homodimers.
In some aspects, the heterodimeric protein of the invention further comprises a first CH2 region and second CH2 region, which interact together to form a CH2 interface, wherein one or more amino acids within the CH2 interface destabilize homodimer formation and are not electrostatically unfavorable to homodimer formation.
In some aspects, the first CH1 or CH2 is connected to a first CH3 domain (CH3), and the second CHI or CH2 is connected to second CH3. The first and second CH3 may each comprise a first and second CH3 mutant residue respectively, the first and second CH3 mutant residues being engineered to differ from each other, and preferentially interact with each other and thereby form CH3 heterodimers preferentially over the formation of CH3 homodimers. Techniques involving replacing one or more residues that make up the CH3-CH3 interface in both CH3 domains with a charged amino acid for promoting the heterodimer formation have also been described in W02009/089004.
In some aspects, the heterodimeric protein of the invention further comprises a

- 54 -first CH3 region and a second CH3 region, which interact together to form a interface, wherein one or more amino acids within the CH3 interface destabilizes homodimer formation and are not electrostatically unfavorable to homodimer formation. In some embodiments, the engineered CH3 interface sterically favors heterodimer formation over homodimer formation. In some embodiments, the engineered CH3 interface electrostatically favors heterodimer formation over homodimer formation.
In some embodiments, the amino acid modification in the first CH3 polypeptide is an amino acid substitution at CH3-391, and the amino acid modification in the second CH3 polypeptide is an amino acid substitution at CH3-441 (according to the numbering of SEQ ID NO:18). In some embodiments, the amino acid modification in the first CH3 polypeptide is CH3-441R and the amino acid modification in the second CH3 polypeptide is CH3-391E or CH3-391D (for greater detail, see W02011/143545).
In some embodiments, the bispecific antibodies further comprise amino acid modification in the first hinge region at positions CH2-D232 and CH2-P241 of SEQ ID
NO: 42 (hinge IgG1), or CH2-C233, CH2-E237, and CH2-P241 of SEQ ID NO: 79 (IgG2 hinge sequence) in one arm, and the substituted/replaced amino acid in the first hinge region has an opposite charge to the corresponding amino acid in the second hinge region in another arm (for greater detail, see W02011/143545).
For example, the amino acid modification in the hinge region can be CH2-D232R, CH2-D232E, CH2-P241 R, and/or CH2-P241 E. In another example, the amino acid modification in the hinge region can be CH2-C233D, CH2-C233E, CH2-C233K, CH2-C223R, CH2-E237E, CH2-E237K, CH2-E237R, CH2-P241D, CH2-P241E, CH2-P241K, and/or CH2-P228R. In some aspects, the CH3 domain is selected from the group consisting of SEQ ID NO:82, 83, 84, and 85.
In some aspects, the invention comprises a CH2 domain comprising a sequence identical to SEQ ID NO:13, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH2 domain comprising a sequence identical to SEQ ID NO:14, by at least an amount selected

- 55 -from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH2 domain comprising a sequence identical to SEQ ID NO:15, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH2 domain comprising a sequence identical to SEQ ID NO:16, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH2 domain comprising a sequence identical to SEQ ID
NO:17, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH2 domain comprising a sequence identical to SEQ ID NO:45, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%.
In some aspects, the invention comprises a CH3 domain comprising a sequence identical to SEQ ID NO:18, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH3 domain comprising a sequence identical to SEQ ID NO:19, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH3 domain comprising a sequence identical to SEQ ID NO:20, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99 /0.In some aspects, the invention comprises a CH3 domain comprising a sequence identical to SEQ ID NO:21, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH3 domain comprising a sequence identical to SEQ ID
NO:22, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some ,

- 56 -aspects, the invention comprises a CH3 domain comprising a sequence identical to SEQ ID NO:23, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH3 domain comprising a sequence identical to SEQ ID NO:46, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH3 domain comprising a sequence identical to SEQ ID NO:47, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH3 domain comprising a sequence identical to SEQ ID NO:48, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%. In some aspects, the invention comprises a CH3 domain comprising a sequence identical to SEQ ID NO:49, by at least an amount selected from the group consisting of 85%, 86, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, and 99%.
In some aspects, the invention further comprises a IgG hinge region between the CHI and CH2 region. The IgG hinge region may comprise SEQ ID NO:42. The IgG hinge region may comprise SEQ ID NO:43. The IgG hinge region may comprise SEQ ID NO:44. The IgG hinge region may be a IgG2 hinge region, and may comprise SEQ ID NO:79.
Ig isotype and subclass In some embodiments, the heterodimeric protein may comprise one or more IgA
domains. In some embodiments, the heterodimeric protein may comprise one or more IgD domains. In some embodiments, the heterodimeric protein may comprise one or more IgE domains. In some embodiments, the heterodimeric protein may comprise one or more IgG domains. In some embodiments, the heterodimeric protein may comprise one or more IgM domains.
In some embodiments, at least one Fab is an IgA1, or IgA2. In some embodiments, at least one Fab is an IgG1, IgG2, IgG3, or IgG4. In some ,

- 57 -embodiments, the IgG Fab comprises a human IgG Fab (e.g. IgGi, IgG2, IgG3, or IgG4). In some embodiments, the first and second Fab are the same subclass (i.e.
both are IgGi, or both are IgG2, or both are IgG3, or both are IgG4).
In alternative embodiments, the first Fab is of a different subclass to the second Fab (i.e. the first Fab and second Fab may each be of a different subclass, and each may be selected from the group consisting of IgGi, IgG2, IgG3, IgG4, IgAi or IgA2).
For example, the antibody of the invention may comprise a first Fab from one antibody sub-class (for example, selected from the group consisting of IgGi, IgG2, IgG3, IgG4, IgAi or IgA2), and a second Fab from a different sub-class (for example, selected from the group consisting of IgGi, IgG2, IgG3, IgG4, IgAi or IgA2, provided the second Fab is of a different subclass to the first Fab), and first and second CH2 domains and first and second CH3 domains from a single antibody class (for example, selected from the group consisting of IgGi, IgG2, IgG3, IgG4, IgAi or IgA2).
In another aspect of the invention, the antibody or Fab region thereof (e.g., bispecific antibody) as described herein comprises a full-length human antibody, wherein a first antibody variable domain of the antibody or Fab thereof is capable of recruiting the activity of a human immune effector cell by specifically binding to an effector antigen located on the human immune effector cell, wherein a second antibody variable domain of the heterodimeric protein is capable of specifically binding to a target antigen. In some embodiments, the human antibody has an IgGi, IgG2, IgG3, or IgG4 isotype.
Except where indicated otherwise by context, the terms "first" and "second", and variations thereof, are merely generic identifiers, and are not to be taken as identifying a specific or a particular CH1, CL, VH, VL, CH2, CH3, or Fab.
In another aspect of the invention, a heterodimeric protein disclosed herein may be deimmunized to reduce immunogenicity upon potential administration to a subject using known techniques such as those described, e.g. in PCT
Publication W098/52976 and W000/34317.
In other embodiments, a heterodimeric protein may be modified or derivatized, such as by making a fusion antibody or immunoadhesin that comprises all or a

- 58 -portion of the heterodimeric polypeptide, e.g. bispecific antibody disclosed herein, linked to another polypeptide or molecular agent. Heteromultimeric, e.g.
heterodimeric polypeptides disclosed herein (e.g., bispecific antibodies) may be modified or derivatized, for example, to potentially extend in vivo half-lives, by producing fusion molecules that may be more stable and/or by treatment with biocompatible polymers such as polyethylene glycol (PEG), commonly referred to as "pegylation," or by any of a number of other engineering methods well known in the art.
A heterodimeric protein may be derivatized with a chemical group, including but not limited to polyethylene glycol (PEG), a methyl or ethyl group, an ester, a carbohydrate group and the like, using well known techniques. These chemical groups (and others like them which have been used to stabilize compounds in vivo) may be useful to improve the biological characteristics of the heterodimeric polypeptide, e.g., to potentially increase serum half-life and bioactivity.
A heterodimeric protein may also be labeled using any of a multitude of methods known in the art. As used herein, the terms "label" or "labeled"
refers to incorporation of another molecule in the antibody. In one embodiment, the label may be a detectable marker, e.g., incorporation of a radio labeled amino acid or attachment to a polypeptide of biotinyl moieties that can be detected by marked avidin (e.g., streptavidin containing a fluorescent marker or enzymatic activity that can be detected by optical or colorimetric methods). In another embodiment, the label or marker may be therapeutic, e.g., a drug conjugate or toxin. Various methods of labeling polypeptides and glycoproteins are known in the art and may be used.
Examples of labels for polypeptides include, but are not limited to:
radioisotopes or radionuclides (e.g., 3H, 14C, 15N, 35S, 90Y, 99Tc, 111In, 1251, 1311), fluorescent labels (e.g., FITC, rhodamine, lanthanide phosphors), enzymatic labels (e.g., horseradish peroxidase, 13-galactosidase, luciferase, alkaline phosphatase), chemiluminescent markers, biotinyl groups, predetermined polypeptide epitopes recognized by a secondary reporter (e.g., leucine zipper pair sequences, binding sites for secondary antibodies, metal binding domains, epitope tags), magnetic ,

- 59 -agents, such as gadolinium chelates, toxins such as pertussis toxin, taxol, cytochalasin B, gramicidin D, ethidium bromide, emetine, mitomycin, etoposide, tenoposide, vincristine, vinblastine, colchicine, doxorubicin, daunorubicin, dihydroxy anthracin dione, mitoxantrone, mithramycin, actinonnycin D, 1-dehydrotestosterone, glucocorticoids, procaine, tetracaine, lidocaine, propranolol, and puromycin and analogs or homologs thereof. In some embodiments, labels may be attached by spacer arms of various lengths to reduce potential steric hindrance.
Nucleic acids and methods of producing polypeptides and heterodimeric proteins of the invention In some embodiments, different nucleic acid molecules encode one or more chains or portions of the heterodimeric protein, e.g. bispecific antibody disclosed herein. In other embodiments, the same nucleic acid molecule encodes a heterodimeric protein disclosed herein.
In one aspect, the present invention provides a nucleic acid sequence encoding one of the chains of a heterodimeric protein disclosed herein, or portion thereof as described above. Nucleic acid molecules of the invention include nucleic acids that hybridize under highly stringent conditions, such as those at least about 70%, 75%, 80%, 85%, 90%, 95%, 97%, 98% or 99% or more identical to a nucleic acid sequence of the invention.
In some aspects, the nucleic acid is DNA. In some aspects, the nucleic acid is RNA. In some aspects, the nucleic acid is mRNA. In some aspects, the nucleic acid is a non-natural nucleic acid, such as PNA (peptide nucleic acid), morpholino and locked nucleic acid, glycol nucleic acid, and threose nucleic acid.
In a further aspect, the present invention provides a vector comprising a nucleic acid sequence encoding one or more of the chains or portions of the heteromultimeric or heterodimeric protein disclosed herein, or portion thereof as described herein.
In a further aspect, the present invention provides a vector suitable for expressing one or more of the chains or portions of the heterodimeric protein disclosed herein, or portion thereof as described herein. In some aspects, the invention provides for a

- 60 -vector that comprises a nucleic acid of the invention.
In another embodiment, a nucleic acid molecule of the invention may be used as a probe or PCR primer for a specific amino acid sequence, e.g. a specific antibody sequence such as in the hinge and constant heavy domain sequences. For instance, the nucleic acid may be used as a probe in diagnostic methods or as a PCR
primer to amplify regions of DNA that might be used, inter alia, to isolate additional nucleic acid molecules encoding useful sequences. In some embodiments, the nucleic acid molecules are oligonucleotides. In some embodiments, the oligonucleotides are from hinge and constant domain regions of the heavy and light chains of an antibody of interest. In some embodiments, the oligonucleotides encode all or a part of one or more of the modified Fab regions of the heterodimeric polypeptide, e.g.
bispecific antibodies or fragments thereof of the invention as described herein.
Recombinant expression vectors of the invention may, in some embodiments, carry regulatory sequences that control the expression of antibody chain genes in a host cell. It will be appreciated by those skilled in the art that the design of the expression vector, including the selection of regulatory sequences may depend on such factors as the choice of the host cell to be transformed, the level of expression of protein desired, etc. Regulatory sequences for mammalian host cell expression may include viral elements that direct high levels of protein expression in mammalian cells, such as promoters and/or enhancers derived from retroviral LTRs, cytomegalovirus (CMV) (such as the CMV promoter/enhancer), Simian Virus 40 (SV40) (such as the SV40 promoter/enhancer), adenovirus, (e.g. the adenovirus major late promoter (AdMLP)), polyoma and strong mammalian promoters such as native immunoglobulin and actin promoters.
In addition to the antibody chain genes and regulatory sequences, the recombinant expression vectors of the invention may carry additional sequences, such as sequences that regulate replication of the vector in host cells (e.g.
origins of replication) and selectable marker genes. For example, the selectable marker gene may confer resistance to drugs, such as G418, hygromycin or methotrexate, on a host cell into which the vector has been introduced. For example, selectable marker

- 61 -genes include the dihydrofolate reductase (DHFR) gene (for use in dhf( host cells with methotrexate selection/amplification), the neo gene (for G418 selection), and the glutamate synthetase gene.
The term "expression control sequence" as used herein means polynucleotide sequences that are necessary to effect the expression and processing of coding sequences to which they are ligated. Expression control sequences include appropriate transcription initiation, termination, promoter and enhancer sequences;
efficient RNA processing signals such as splicing and polyadenylation signals;

sequences that stabilize cytoplasmic mRNA; sequences that enhance translation efficiency (i.e., Kozak consensus sequence); sequences that enhance protein stability; and when desired, sequences that enhance protein secretion. The nature of such control sequences differs depending upon the host organism; in prokaryotes, such control sequences generally include promoter, ribosomal binding site, and transcription termination sequence; in eukaryotes, generally, such control sequences include promoters and transcription termination sequence. The term "control sequences" is intended to include, at a minimum, all components whose presence is essential for expression and processing, and may also include additional components whose presence may be advantageous, for example, leader sequences and fusion partner sequences.
In some aspects, the invention comprises a nucleic acid encoding at least one CH1, or CL of the invention. The invention further provides for nucleic acids that encode for a Fab of the invention. In some aspects, the invention provides for a nucleic acid that encodes for a first Fab of the invention. In some aspects, the invention provides for a nucleic acid that encodes for a second Fab of the invention.
In some aspects, the invention provides for a nucleic acid that encodes a first heavy chain of the invention. In some aspects, the invention provides for a nucleic acid that encodes a second heavy chain of the invention. In some aspects, the invention provides for a nucleic acid that encodes a first light chain of the invention. In some aspects, the invention provides for a nucleic acid that encodes a second light chain of the invention.

- 62 -In some aspects, the invention provides for a cell that comprises a vector of the invention. In some aspects, the invention provides for a cell that comprises a nucleic acid of the invention. In some aspects, the invention provides for a cell that expresses a nucleic acid of the invention.
In one aspect, the invention provides for a cell that expresses a heterodimeric protein as herein described. Co-expressing the first CHCL and second CHCL in the same cell takes advantage of the complementary residue sets that allow for correct formation of the heteromultimeric protein. In some aspects, this may permit a bispecific antibody to be expressed and generated in a fully assembled form, and may require little to no additional purification or processing steps over what would be typically required for purification of a monoclonal antibody.
In some aspects, bispecific antibodies of the invention may potentially be used in mRNA replacement therapy or may potentially be used in RNA transcript therapy.

Accordingly, in some aspects, the invention comprises a cell, or vector, comprising one or more nucleic acids encoding one or more polypeptide chains of the invention, such that expression of the polypeptide chains of the invention in a potential in vivo application results in the generation of a bispecific antibody in vivo.
Delivery mechanisms for such vectors include lipid based systems and nanoparticles (see for example, W02010053572, W02012170930 and W02011068810, each of whose contents is incorporated entirely).
In some aspects, the invention further comprises a transfer vehicle, defined herein as any of the standard pharmaceutical carriers, diluents, excipients and the like which may be used in connection with the potential administration of biologically active agents, including nucleic acids. The compositions and in particular the transfer vehicles described herein may be used for delivering nucleic acids of the invention to a target cell. In some embodiments, the transfer vehicle is a lipid nanoparticles, which may be used for transferring mRNA to a target cell.
In some aspects, the invention comprises an mRNA encoding a bispecific antibody of the invention, a transfer vehicle and, optionally, an agent to facilitate contact with, and subsequent transfection of a potential target cell.

- 63 -In some embodiments the mRNA encoding one or more polypeptides of the invention can comprise one or more modifications that may confer stability to the mRNA (e.g., compared to a wild-type or native version of the mRNA). For example, the nucleic acids of the present invention may comprise modifications to one or both of the 5' and 3' untranslated regions. Such modifications may include, but are not limited to, the inclusion of a partial sequence of a cytomegalovirus (CMV) immediate-early 1 (1E1) gene, a poly A tail, a Capl structure or a sequence encoding human growth hormone (hGH)). In some embodiments, the mRNA may be modified to potentially decrease mRNA immunogenicity.
In some embodiments, the mRNA of the invention have undergone a chemical or biological modification to render them potentially more stable. Exemplary modifications to an mRNA may include the depletion of a base (e.g. by deletion or by the substitution of one nucleotide for another) or modification of a base, for example, the chemical modification of a base. In some aspects, a poly A tail may be added to an mRNA molecule thus rendering the mRNA potentially more stable.
In some aspects, the transfer vehicle in the compositions of the invention is a liposomal transfer vehicle, e.g. a lipid nanoparticle. The transfer vehicle may be selected and/or prepared to potentially optimize delivery of the mRNA to a target cell.
For example, if the target cell is a hepatocyte the properties of the transfer vehicle (e.g., size, charge and/or pH) may be optimized to potentially effectively deliver such transfer vehicle to the target cell, potentially reduce immune clearance and/or potentially promote retention in that target cell. Alternatively, if the target cell is the central nervous system (e.g. mRNA administered for the potential treatment of neurodegenerative diseases may specifically target brain or spinal tissue), selection and preparation of the transfer vehicle must consider penetration of, and retention within the blood brain barrier and/or the use of alternate means of directly delivering such transfer vehicle to such potential target cell. In some aspects, the compositions of the present invention may be combined with agents that facilitate the transfer of exogenous mRNA (e.g. agents which may disrupt or improve the permeability of the blood brain barrier and thereby potentially enhance the transfer of exogenous mRNA

- 64 -to the target cells).
The potential use of liposomal transfer vehicles to facilitate the delivery of nucleic acids to target cells is contemplated by the present invention. In some aspects, the transfer vehicle may be formulated as a lipid nanoparticle. Examples of lipids may include, for example, the phosphatidyl compounds (e.g., phosphatidylglycerol, phosphatidylcholine, phosphatidylserine, phosphatidylethanolamine, sphingolipids, cerebrosides, and gangliosides). Also contemplated is the potential use of polymers as transfer vehicles, whether alone or in combination with other transfer vehicles.
Examples of polymers may include, for example, polyacrylates, polyalkycyanoacrylates, polylactide, polylactide- polyglycolide copolymers, polycaprolactones, dextran, albumin, gelatin, alginate, collagen, chitosan, cyclodextrins, dendrimers and polyethylenimine.
The invention contemplates the potential use of lipid nanoparticles as transfer vehicles comprising a cationic lipid that may encapsulate and/or enhance the delivery of mRNA into the target cell that may act as a depot for protein production.
As used herein, the phrase "cationic lipid" refers to any of a number of lipid species that can a net positive charge at a selected pH, such as physiological pH.
In one aspect, this invention provides a strategy for enhancing the formation of a bispecific antibody, by altering or engineering an interface between the light chain and the heavy chain of one or more Fab regions of the antibody. In some embodiments, one or more residues that make up the CH1/CL interface of the one more Fab regions are replaced with residues such that the modified residues favor pairing of the specific heavy and light chain of the modified Fab region over mispairing with heavy chains or light chains of other Fab regions in the protein. In one embodiment, the modifications introduce novel disulfide bridges in the Fab region. In another embodiment, the modifications introduce disrupting mutations that disrupt the native interface between the CH1 and CL domains of a Fab region, as well as restoring modifications that introduce non-native stable interactions at the interface.
In another embodiment, the disrupting mutations may introduce both novel disulfide bridges and disrupting and restoring mutations.

- 65 -In some embodiments, the formation of the heterodimeric protein comprising one or more amino acid modifications in the CH1/CL interface of one or more Fab regions disclosed herein may be substantially increased in comparison to the wild-type heterodimeric protein without such modifications. In some embodiments, the formation of the heterodimeric protein comprising one or more amino acid modifications in CH1/CL interface of at least one Fab region may be at least about any of 51%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, 97%, 99%, or 100% in comparison to the wild-type heterodimeric protein without such modifications.
In another aspect, the present invention also provides methods of producing a heteromultimeric protein, e.g. a heterodimeric protein of the invention, such as a bispecific antibody. In some embodiments, the method comprises the steps of:
a) cotransfecting a cell line with vectors expressing each heavy chain and each light chain of each Fab region of the protein;
b) culturing the cell line under conditions to express each heavy chain and each light chain of each Fab region of the protein and that allow the heteromultimeric protein to assemble; and c) purifying the heteromultimeric protein from the cell culture. In some embodiments, the cell line is cotransfected with vectors that express the heavy chain and the light chain of each Fab region in a 1:1:1:1 ratio.
In some embodiments, the method comprises the steps of:
(i) cotransfecting a cell line with one or more vectors to express the first CH1, the first CL of the first CHCL; and the second CH1, and the second CL of the second CHCL;
(i) culturing the cell line under conditions to express the one or more vectors and that allow the first CHCL and second CHCL to assemble; and (ii) purifying the heteromultimeric protein from the cell culture.
In some aspects, the cell line is cotransfected with vectors that express the first CH1, first CL, second CH1, and second CL in a 1:1:1:1 ratio.
The skilled artisan can readily determine, using well-known techniques, the relative amounts of molecules or antibodies to use according to the methods

- 66 -disclosed herein.
In the methods disclosed herein, incubations may be performed across a range of temperatures. Such temperatures will be recognized by those skilled in the art and will include, for example, incubation temperatures at which deleterious physical changes such as denaturation or decomposition do not occur in the mixed molecules or antibodies. In certain embodiments, the incubations are performed at about 37 C.
Any of a number of host cells may be used in methods of the invention. Such cells are known in the art (some of which are described herein) or can be determined empirically with respect to suitability for use in methods of the invention using routine techniques known in the art. In certain embodiments, the host cell is prokaryotic. In some embodiments, a host cell is a gram-negative bacteria cell. In other embodiments, a host cell is E. coil. In some embodiments, the E. coil is of a strain deficient in endogenous protease activities. In some embodiments, the genotype of an E. coil host cell lacks degP and prc genes and harbors a mutant spr gene.
In other embodiments of the invention, the host cell is mammalian, for example, a Chinese Hamster Ovary (CHO) cell.
In some embodiments, methods of the invention further comprise expressing in a host cell a polynucleotide or recombinant vector encoding a molecule the expression of which in the host cell enhances yield of a bispecific antibody or a heterodimeric protein of the invention. For example, such molecule can be a chaperone protein. In one embodiment, said molecule is a prokaryotic polypeptide selected from the group consisting of DsbA, DsbC, DsbG and FkpA. In some embodiments of these methods, the polynucleotide encodes both DsbA and DsbC.
In one aspect, the present invention provides recombinant host cells allowing the recombinant expression of the antibodies of the invention or portions thereof.
Antibodies produced by such recombinant expression in such recombinant host cells are referred to herein as "recombinant antibodies". The present invention also provides progeny cells of such host cells, and antibodies produced by same.
The term "recombinant host cell" (or simply "host cell"), as used herein, means a cell into

- 67 -which a recombinant expression vector has been introduced. It should be understood that "recombinant host cell" and "host cell" mean not only the particular subject cell but also the progeny of such a cell. Because certain modifications may occur in succeeding generations due to either mutation or environmental influences, such progeny may not, in fact, be identical to the parent cell, but are still included within the scope of the term "host cell" as used herein. Such cell may comprise a vector according to the invention as described above.
In another aspect, the present invention provides a method for making an antibody or portion thereof as described above. According to one embodiment, said method comprises culturing a cell transfected or transformed with a vector as described above, and retrieving said antibody or portion thereof. Nucleic acid molecules encoding antibodies and vectors comprising these nucleic acid molecules may be used for transfection of a suitable mammalian, plant, bacterial or yeast host cell. Transformation may be by any known method for introducing polynucleotides into a host cell. Methods for introduction of heterologous polynucleotides into mammalian cells are well known in the art and include dextran-mediated transfection, calcium phosphate precipitation, polybrene-mediated transfection, protoplast fusion, electroporation, encapsulation of the polynucleotide(s) in liposomes, and direct microinjection of the DNA into nuclei. In addition, nucleic acid molecules may be introduced into mammalian cells by viral vectors. Methods of transforming cells are well known in the art. See, e.g., U.S. Patent Nos. 4,399,216, 4,912,040, 4,740,461, and 4,959,455. Methods of transforming plant cells are well known in the art, including, e.g., Agrobacterium-mediated transformation, biolistic transformation, direct injection, electroporation and viral transformation. Methods of transforming bacterial and yeast cells are also well known in the art.
Mammalian cell lines available as hosts for expression are well known in the art and include many immortalized cell lines available from the American Type Culture Collection (ATCC). These include, inter alia, Chinese hamster ovary (CHO) cells, NSO cells, SP2 cells, HEK-293T cells, 293 Freestyle cells (Invitrogen), cells, HeLa cells, baby hamster kidney (BHK) cells, African green monkey kidney

- 68 -cells (COS), human hepatocellular carcinoma cells (e.g., Hep G2), A549 cells, and a number of other cell lines. Cell lines may be selected through determining which cell lines have high expression levels. Other cell lines that may be used are insect cell lines, such as Sf9 or Sf21 cells. When recombinant expression vectors encoding antibody genes are introduced into mammalian host cells, the antibodies are produced by culturing the host cells for a period of time sufficient to allow for expression of the antibody in the host cells or, more preferably, secretion of the antibody into the culture medium in which the host cells are grown. Antibodies can be recovered from the culture medium using standard protein purification methods.
Suitable plant host cells may include, e.g., Nicotiana, Arabidopsis, duckweed, corn, wheat, potato, etc. Suitable bacterial host cells may include, e.g., E. coil and Streptomyces species. Suitable yeast host cells may include, e.g., Schizosaccharomyces pombe, Saccharomyces cerevisiae and Pichia pastoris.
Expression of polypeptides of the invention or portions thereof from production cell lines may be enhanced using a number of known techniques. For example, the glutamine synthetase gene expression system (the GS system) is a common approach for enhancing expression under certain conditions. The GS system is discussed in whole or part in connection with EP0216846, EP0256055, EP0323997, and EP0338841.
It is likely that polypeptides comprising Fc polypeptides or Fc regions and immunoglobulin-like hinge polypeptides, such as antibodies, as expressed by different cell lines or in transgenic animals, will differ from each other in their glycosylation patterns. All such "glycoforms" of polypeptides of the invention, including all heterodimers of polypeptides comprising immunoglobulin-like hinge sequences, bispecific polypeptides, antibodies and the like, are considered to be part of the instant invention, regardless of their glycosylation state, and more generally, regardless of the presence or absence of any post-translational modification(s).
In some embodiments, heterodimeric protein is an antibody, a maxibody, a monobody, a peptibody, an Fc fusion protein, or Fab region of any of the foregoing. In some embodiments, the heterodimeric protein is a bispecific antibody.

- 69 -The heterodimeric protein thereof may comprise one or more human domains.
The heterodimeric protein may comprise one or more humanized Ig domains. The heterodimeric protein may comprise one or more murine Ig domains. The heterodimeric protein may comprise one or more Ig domains originating from a species selected from the group consisting of human, monkey, mouse, rat, hamster, guinea pig, rabbit, dog, cat, donkey, goat, camel, cow, horse, pig, chicken, and shark.
In some aspects, the antibodies of the invention are mammalian, avian, or Squaliform in origin (notwithstanding the method used to generate any artificially mutated or otherwise engineered versions). The mammalian, avian, or squaliform species may be human, mouse, rabbit, rat, rodent, pig, cow, sheep, goat, donkey, horse, camel, llama, primate, monkey, dog, cat, chicken, or spiny dogfish. The antibodies of the invention may be humanized.
In some aspects, the invention comprises mutant antibodies and portions thereof, wherein a mutant is defined as sequence that has been engineered or altered to a sequence other than its natural canonical sequence, such that certain embodiments of polypeptides of the invention specifically excludes naturally occurring sequences that fall within the scope of the definition. In some aspects, therefore, the present invention relates to polypeptides of the invention comprising mutations to enable heterodimeric Ig-domain pairing such that the Ig domain polypeptide sequence differs from its naturally occurring corresponding sequence.
Antibody CH1 domains may be selected from the group consisting of CHa1, CH61, CHE1, CHy1, and CHp1.
In some aspects, the constant light chain (CL) domain of the invention is connected to a variable light chain (VL) domain. Together, these may comprise an antibody light chain. The CL domain may be a CLK (constant light chain kappa).
The CL domain may be a CLA (constant light chain lambda).
In some aspects, the CHI domain of the invention is connected to a variable heavy chain (VH) domain. Together, these may comprise the heavy chain portion of a Fab molecule. In some aspects, the VH and CH1 domains are connected to the remainder of the CH domains typical for that particular Ig isotpye (i.e. CHa1 may be

- 70 -connected to CHa2, and CHa3; 0H61 may be connected to CH62 and CH63; CHE1 may be connected to CHE2, CH63, and CHE4; CHy1 may be connected to CHy2, and CHY3;

CHp1may be connected to CHp2, CHp3, and CHp4).
In some aspects, the invention provides for an isolated host cell that recombinantly produces an antibody of the present invention. The present invention provides for an isolated polynucleotide comprising a nucleotide sequence encoding proteins, domains and antibodies of the present invention, and vectors comprising said polynucleotides. In some aspects, the invention provides for a method of producing an antibody, immunoglobulin domain, or protein, comprising culturing a host cell under conditions that result in production of the antibody, immunoglobulin domain, or protein, and isolating the antibody, immunoglobulin domain, or protein, from the host cell or culture.
The invention provides improved methods, compositions, kits and articles of manufacture for generating heteromultimeric complex molecules, more preferably, heterodimeric proteins, such as, e.g. a bispecific antibody. The invention may provide methods to make and to purify heteromultimeric complex molecules in yields and purities desirable for commercial manufacture of potential biotherapeutics.
The invention makes possible efficient production of complex molecules that, in turn, may potentially be used for diagnosing and/or treating various disorders or conditions where use of multispecific antibodies is desirable and/or required.
Pharmaceutical compositions A pharmaceutical composition comprising proteins of the invention of the invention and a pharmaceutically acceptable carrier may also be prepared. As used herein, "pharmaceutically acceptable carrier" includes any and all solvents, dispersion media, coatings, antibacterial and antifungal agents, isotonic and absorption delaying agents, and the like that are physiologically compatible.
In certain embodiments, the proteins of the invention may be present in a neutral form (including zwitter ionic forms) or as a positively or negatively-charged species. In some embodiments, the polypeptides may be complexed with a counterion to form a "pharmaceutically acceptable salt," which refers to a complex comprising one or more

- 71 -polypeptides and one or more counterions, where the counterions are derived from pharmaceutically acceptable inorganic and organic acids and bases.
Antibodies An "Antibody" is an immunoglobulin molecule capable of specific binding to a target or antigen, such as a carbohydrate, polynucleotide, lipid, polypeptide, etc., through at least one antigen-binding site, located in the variable region of the immunoglobulin molecule.
As used herein, unless otherwise indicated by context, the term is intended to encompass not only intact polyclonal or monoclonal antibodies comprising two identical full-length heavy chain polypeptides and two identical light chain polypeptides, but also fragments thereof (such as Fab, Fab', F(ab')2, Fv), single chain (ScFv) and domain antibodies (dAbs), including shark and camelid antibodies, and fusion proteins comprising an antibody portion, multivalent antibodies, multispecific antibodies (e.g. bispecific antibodies so long as they exhibit the desired biological activity) and antibody fragments as described herein, and any other modified configuration of the immunoglobulin molecule that comprises an antigen recognition site, for example without limitation, minibodies, maxibodies, monobodies, peptibodies, intrabodies, diabodies, triabodies, tetrabodies, v-NAR and bis-scFv.
Antigen-binding portions may be produced by recombinant DNA techniques or by enzymatic or chemical cleavage of intact antibodies. Antigen-binding portions include, inter alia, Fab, Fab', F(ab')2, Fv, dAb, and complementarity determining region (CDR) fragments, single-chain antibodies (scFv), chimeric antibodies, diabodies and polypeptides that contain at least a portion of an Ig that is sufficient to confer specific antigen binding to the polypeptide.
The immunoglobulin (Ig) domain is a type of protein domain that typically consists of a 2-layer sandwich of between 7 and 9 (3-strands arranged in two i3-sheets (although variations on these arrangements are known). A 13-strand is a stretch of polypeptide chain typically 3 to 10 amino acids long with backbone in an almost fully extended conformation. 13 sheets consist of 13-strands connected laterally by at least two or three backbone hydrogen bonds, forming a generally twisted,

- 72 -pleated sheet. The backbone of a strand switches repeatedly between interacting with its two opposite neighboring strands in the sheet, or between sheet and non-sheet interactions for strands at the sheet edge. Members of the Ig superfamily are found in hundreds of proteins of different functions. Examples include antibodies, the giant muscle kinase titin and receptor tyrosine kinases. Ig-like domains may be involved in protein¨protein and protein¨ligand interactions.
An immunoglobulin (Ig) is a heteromultimeric molecule. In a naturally occurring Ig, each multimer is composed primarily of identical pairs of polypeptide chains, each pair having one "light" (about 25 kDa) and one "heavy" chain (about 50-70 kDa).
The amino-terminal portion of each chain includes a variable region, of about 100 to 110 or more amino acids primarily responsible for antigen recognition.
The carboxy-terminal portion of each chain defines a constant region primarily responsible for effector function. Human light chains are classified as K and A light chains. Heavy chains are classified as a, 6, 6, y, and p, and define the antibody's isotype as IgA, IgD, IgE, IgG, IgM, respectively. Several of these classes may be further subdivided into isotypes: IgG1, IgG2, IgG3, IgG4, IgA1, and IgA2.
Within light and heavy chains, the variable and constant regions are joined by a "J" region of about 12 or more amino acids, with the heavy chain also including a "D" region of about 10 more amino acids (in the context of an entire antibody sequence, the D and J regions are sometimes considered as parts of the variable region after they have been joined). The variable regions of each light/heavy chain pair form the antibody binding site such that an intact Ig has 2 binding sites.
Each domain in an antibody molecule has a similar structure of two 13-sheets packed tightly against each other in a compressed antiparallel 13-barrel. This conserved structure is termed the immunoglobulin (Ig) fold. The Ig fold of constant domains contains two 13 sheets packed against each other, with each strand separated by a contiguous polypeptide string; these contiguous polypeptide strings typically comprise a-helices, loops, turns, and short, sharp turns between two sheets called 13-hairpins.
Variable domains exhibit the same general structure of relatively conserved

- 73 -framework regions (FR) joined by 3 hypervariable regions, also called complementarity determining regions or CDRs. The CDRs from the 2 chains of each pair are aligned by the framework regions, enabling binding to a specific epitope.
From N-terminus to C-terminus, both light and heavy chains comprise the domains FR1, CDR1, FR2, CDR2, FR3, CDR3 and FR4.
The identity of the amino acid residues in a particular antibody that make up a CDR can be determined using methods well known in the art. For example, antibody CDRs may be identified as the hypervariable regions originally defined by Kabat et al (Kabat et a/.,1991, Sequences of Proteins of Immunological Interest, 5th ed., Public Health Service, NIH, Washington D.C., NIH Publication No. 91-3242). The positions of the CDRs may also be identified as the structural loop structures described by Chothia and others (Chothia et al., 1989, Nature 342:877-883). Other approaches to CDR identification include the "AbM definition," which is a compromise between Kabat and Chothia and is derived the Abysis program (www.abysis.org), or the "contact definition" of CDRs based on observed antigen contacts, set forth in MacCallum et al., 1996, J. Mol. Biol., 262:732-745. North has identified canonical CDR conformations using a different preferred set of CDR definitions (North et at., 2011, J. Mol. Biol, 406: 228-256). In another approach, referred to herein as the "conformational definition" of CDRs, the positions of the CDRs may be identified as the residues that make enthalpic contributions to antigen binding (Makabe et at., 2008, Journal of Biological Chemistry, 283:1156-1166). Still other CDR
boundary definitions may not strictly follow one of the above approaches, but will nonetheless overlap with at least a portion of the Kabat CDRs, although they may be shortened or lengthened in light of prediction or experimental findings that particular residues or groups of residues or even entire CDRs do not significantly impact antigen binding.
As used herein, a CDR may refer to CDRs defined by any approach known in the art, including combinations of approaches. The methods used herein may utilize CDRs defined according to any of these approaches. For any given embodiment containing more than one CDR, the CDRs (or other residue of the antibody) may be defined in accordance with any of Kabat, Chothia, North, extended, AbM, contact, and/or

- 74 -conformational definitions.
Except where indicated otherwise explicitly or by context, all CH1 residue numbering positions herein described are according to the numbering of SEQ ID
NO:1, and all CL residue positions are herein described according to the numbering of SEQ ID NO:9. This numbering is most closely related to the numbering of Kabat, which is used herein except (a) in cases such as IgM domain where certain experimental data has shown Kabat to be incorrect, (b) when Kabat's reference is internally inconsistent, or (c) when otherwise noted. In the original Kabat reference, position 107A is the first residue of the CL. Many light chain sequences do not have any residue assigned to position 107A and many also do not have a residue at position 108. The first residue of CL is the first residue numbered greater than 107, whatever that may be.
A CH1 domain is a region of protein sequence, preferably at least 80 residues in length, and having more than 85% of its residues in common with one or more of SEQ ID NO:1, SEQ ID NO:2, SEQ ID NO:3, SEQ ID NO:4, SEQ ID NO:5, SEQ ID
NO:6, SEQ ID NO:7, SEQ ID NO:8, SEQ ID NO:33, SEQ ID NO:34, SEQ ID NO:35, SEQ ID NO:36, SEQ ID NO:37, SEQ ID NO:38, SEQ ID NO:39, SEQ ID NO:40, or SEQ ID NO:41. In some aspects, a CH1 domain is protein sequence having more than 85% of its residues in common with SEQ ID NO:1.
A CH2 domain is a region of protein sequence, preferably at least 80 residues in length, and having more than 85% of its residues in common with one or more of SEQ ID NO:13, SEQ ID NO:14, SEQ ID NO:15, SEQ ID NO:16, SEQ ID NO:17, or SEQ ID NO:45. In some aspects, a CH2 domain is protein sequence having more than 85% of its residues in common with SEQ ID NO:13.
A hinge region is a region of protein sequence having more than 80% identity with one or more of SEQ ID NO:42, SEQ ID NO:43, SEQ ID NO:44, SEQ ID NO:79, SEQ ID NO:80, or SEQ ID NO:81. In some aspects, a hinge region is protein sequence having more than 80% of its residues in common with SEQ ID NO:42.
A CH3 domain is a region of protein sequence, preferably at least 80 residues in length, and having more than 85% of its residues in common with one or more of

- 75 -SEQ ID NO:18, SEQ ID NO:19, SEQ ID NO:20, SEQ ID NO:21, SEQ ID NO:22, SEQ
ID NO:23, SEQ ID NO:46, SEQ ID NO:47, SEQ ID NO:48, or SEQ ID NO:49. In some aspects, a CH3 domain is protein sequence having more than 85% of its residues in common with SEQ ID NO:18.
A CL domain is a region of protein sequence preferably at least 80 residues in length, and having more than 85% of its residues in common with one or more of SEQ ID NO:9, SEQ ID NO:10, SEQ ID NO:11, SEQ ID NO:12, SEQ ID NO:24, SEQ
ID NO:25, SEQ ID NO:26, SEQ ID NO:27, SEQ ID NO:28, SEQ ID NO:29, SEQ ID
NO:30, SEQ ID NO:31, or SEQ ID NO:32. In some aspects, the CL domain is a CL
kappa domain, and shares at least 85% identity with one or more of SEQ ID
NO:9, SEQ ID NO:10, or SEQ ID NO:11. In some aspects, the CL domain is a CL lambda domain, and shares at least 85% identity with SEQ ID NO:12. In some aspects, a CL
domain is protein sequence having more than 85% of its residues in common with SEQ ID NO:9.
Mammalian light chains are of two types, K and A, and in any given naturally occurring antibody molecule only one type occurs. Approximately twice as many K as A molecules are produced in humans but in other mammals this ratio can vary.
Each free light chain molecule contains approximately 220 amino acids in a single polypeptide chain that is folded to form the constant and variable region domains.
During B cell development, a recombination event at the DNA level joins a single variable (V) segment with a joining (J) segment; the constant (C) segment is later joined by splicing at the RNA level. Recombination of many different V
segments with several J segments provides a wide range of antigen recognition.
Additional diversity is attained by junctional diversity, resulting from the random additional of nucleotides by terminal deoxynucleotidyltransferase, and by somatic hypermutation, which occurs during B cell maturation in the spleen and lymph nodes. Constant kappa (CLK) regions are encoded by a single gene, whereas lambda constant (CLA) regions are encoded by multiple genes, and undergo splicing. Several markers associated with particular polymorphic species of CLA are known: IgCLA1 (Mcg marker); IgLC2 ¨ IgCUk2 (Kern-Oz- marker); IgCLA 3 (Kern-Oz+ marker), and

- 76 -1gCLA7, for example. The skilled person can easily establish all of the polymorphisms so far identified in human CLA chains. The sequences of the present invention encompass other known polymorphisms of the CLK and CLA, and antibodies in general. Two polymorphic loci have been identified in the CLK; CLK-V/A153 and CLK-LN191. The three polymorphisms so far identified are: Km(1): CLK-V153/L191;
Km(1,2):
CLK-A153/ L191;
and Km(3): CLK-A153N191.
The term "Fc region" as used herein generally refers to a dimer complex comprising the C-terminal polypeptide sequences of an immunoglobulin heavy chain, wherein a C-terminal polypeptide sequence is that which is obtainable by papain digestion of an intact antibody. The Fc region may comprise native or variant Fc sequences. The Fc sequence of an immunoglobulin generally comprises two constant domains, a CH2 domain and a CH3 domain, and optionally comprises a domain. The term "Fc polypeptide" is used herein to refer to one of the polypeptides that makes up an Fc region. In some embodiments, an Fc polypeptide may be obtained or derived from any suitable immunoglobulin, such as from at least one of the various IgG1, IgG2, IgG3, or IgG4 subtypes, or from IgA, IgE, IgD or IgM.
In some embodiments, an Fc polypeptide comprises part or all of a wild-type hinge sequence (generally at its N terminus). In some embodiments, an Fc polypeptide does not comprise a wild-type hinge sequence. An Fc polypeptide may comprise native or variant Fc sequences.
The "immunoglobulin-like hinge region," "immunoglobulin-like hinge sequence," and variations thereof, as used herein, refer to the hinge region and hinge sequence of an immunoglobulin-like or an antibody-like molecule (e.g.
immunoadhesins). In some embodiments, the immunoglobulin-like hinge region can be from or derived from any IgG1, IgG2, IgG3, or IgG4 subtype, or from IgA, IgE, IgD
or IgM, including chimeric forms thereof, e.g. a chimeric IgG1/2 hinge region.

"Antibody fragments" comprise only a portion of an intact antibody, wherein the portion preferably retains at least one, preferably most or all, of the functions normally associated with that portion when present in an intact antibody.
A "bivalent antibody" comprises two antigen binding sites per molecule (e.g.

- 77 -IgG). In some instances, the two binding sites have the same antigen specificities.
However, bivalent antibodies may be bispecific (see below).
A "monovalent antibody" comprises one antigen binding site per molecule (e.g.
IgG). In some instances, a monovalent antibody can have more than one antigen binding site, but the binding sites are from different antigens.
A "multispecific antibody" is one that targets more than one antigen or epitope.
A "bispecific," "dual-specific" or "bifunctional" antibody is a hybrid antibody having two different antigen binding sites. Bispecific antibodies are a species of multispecific antibody and may be produced by a variety of methods including, but not limited to, fusion of hybridomas or linking of Fab' fragments. See, e.g. Songsivilai &
Lachmann (1990), Clin. Exp. Immunol. 79:315-321; and Kostelny et al. (1992), J.
Immunol.
148:1547-1553. The two binding sites of a bispecific antibody will bind to two different epitopes, which may reside on the same or different protein targets.
The phrase "antigen binding arm," "target molecule binding arm," and variations thereof, as used herein, refers to a component part of an antibody of the invention that has an ability to specifically bind a target molecule of interest.
Generally and preferably, the antigen binding arm is a complex of immunoglobulin polypeptide sequences, e.g. CDR and/or variable domain sequences of an immunoglobulin light and heavy chain.
The term "monoclonal antibody" as used herein refers to an antibody obtained from a population of substantially homogeneous antibodies, i.e., the individual antibodies comprising the population are identical except for possible naturally occurring mutations that may be present in minor amounts. Monoclonal antibodies are highly specific, being directed against a single antigen. Further, in contrast to polyclonal antibody preparations that typically include different antibodies directed against different determinants (epitopes), each monoclonal antibody is directed against a single determinant on the antigen.
As used herein, the term "immunoadhesin" designates antibody-like or immunoglobulin-like molecules which combine the "binding domain" of a heterologous protein (an "adhesin", e.g. a receptor, ligand or enzyme) with the

- 78 -effector component of immunoglobulin constant domains. Structurally, the immunoadhesins comprise a fusion of the adhesin amino acid sequence with the desired binding specificity which is other than the antigen recognition and binding site (antigen combining site) of an antibody (i.e. is "heterologous") and an immunoglobulin constant domain sequence. The immunoglobulin constant domain sequence in the immunoadhesin may be obtained from any immunoglobulin, such as IgG1 , IgG2, IgG3, or IgG4 subtypes, IgA, IgE, IgD or IgM.
A Fab fragment is a monovalent fragment consisting of the VL, VH, CL and CH1 domains; a F(ab')2 fragment is a bivalent fragment comprising two Fab fragments linked by a disulfide bridge at the hinge region; a Ed fragment consists of the VH and CH1 domains; an Fv fragment consists of the VL and VH domains of a single arm of an antibody; and a dAb fragment consists of a VH domain or a VL domain (e.g.
human, camelid, or shark).
A single-chain antibody (scFv) is an antibody in which a VL and VH region are paired to form a monovalent molecule via a synthetic linker that enables them to be made as a single protein chain. Diabodies are bivalent, bispecific antibodies in which VH and VL domains are expressed on a single polypeptide chain, but using a linker that is too short to allow for pairing between the 2 domains on the same chain, thereby forcing the domains to pair with complementary domains of another chain and creating 2 antigen binding sites. One or more CDRs may be incorporated into a molecule either covalently or noncovalently to make it an immunoadhesin. An immunoadhesin may incorporate the CDR (s) as part of a larger polypeptide chain, may covalently link the CDR (s) to another polypeptide chain, or may incorporate the CDR(s) noncovalently. The CDRs permit the immunoadhesin to specifically bind to a particular antigen of interest.
An antibody may have one or more binding sites. If there is more than one binding site, the binding sites may be identical to one another or may be different. For instance, a naturally-occurring antibody has 2 identical binding sites, a single-chain antibody or Fab fragment has one binding site, while a"bispecific"or "bifunctional"
antibody has 2 different binding sites.

- 79 -An "isolated antibody" is an antibody that (1) is not associated with naturally-associated components, including other naturally-associated antibodies, that accompany it in its native state, (2) is free of other proteins from the same species, (3) is expressed by a cell that does not naturally express the antibody, or is expressed by a cell from a different species, or (4) does not occur in nature.
The term "human antibody" includes all antibodies that have one or more variable and constant regions derived from human Ig sequences. In some embodiments of the present invention, all of the variable and constant domains of the antibody are derived from human Ig sequences (a fully human antibody).
A humanized antibody is an antibody that is derived from a non-human species, in which certain amino acids have been mutated so as to avoid or abrogate an immune response in humans. Alternatively, a humanized antibody may be produced by fusing the constant domains from a human antibody to the variable domains of a non-human species.
The term "chimeric antibody" refers to an antibody that contains one or more regions from one antibody and one or more regions from one or more other antibodies. Each antibody may originate from seperate species (such as human and mouse).
The term "epitope" includes any molecular determinant capable of specific binding to an Ig or T-cell receptor. Epitopic determinants usually consist of surface groupings of atoms such as amino acids or sugar side chains and usually have specific 3 dimensional structural characteristics, as well as specific charge characteristics. An antibody is said to specifically bind an antigen when the dissociation constant is <1uM, preferably <100nM and more preferably: <10nM.
Fully human antibodies may be expected to minimize the immunogenic and allergic responses intrinsic to mouse or mouse-derivatized monoclonal antibodies (Mabs) and thus to potentially increase the efficacy and safety of the administered antibodies.
In addition, fusion antibodies can be created in which 2 (or more) single-chain antibodies are linked to one another. This may be useful if one wants to create a

- 80 -divalent or polyvalent antibody on a single polypeptide chain, or if one wants to create a bispecific antibody.
By "Fc fusion" as used herein is meant a protein wherein one or more polypeptides is operably linked to an Fc polypeptide. An Fc fusion combines the Fe region of an immunoglobulin with a fusion partner, which in general may be any protein, polypeptide or small molecule. Virtually any protein or small molecule may be linked to Fc to generate an Fc fusion. Protein fusion partners may include, but are not limited to, the target-binding region of a receptor, an adhesion molecule, a ligand, an enzyme, a cytokine, a chemokine, or some other protein or protein domain.
Small molecule fusion partners may include an agent that directs the Fc fusion to a potential therapeutic target. Such targets may be any molecule, for example without limitation, an extracellular receptor that may be implicated in a disease.
One type of derivatized antibody is produced by erosslinking 2 or more antibodies (of the same type or of different types; e. g. to create bispecific antibodies). Suitable crosslinkers include those that are heterobifunctional, having 2 distinctly reactive groups separated by an appropriate spacer (e. g., m-maleimidobenzoyl-N-hydroxysuccinimide ester) or homobifunctional (e. g.
disuccinimidyl suberate).
Another type of derivatized antibody is a labelled antibody. Useful detection agents with which an antibody or antibody portion of the invention may be derivatized include fluorescent compounds, including fluorescein, fluorescein isothiocyanate, rhodamine, 5-dimethylamine-1-napthalenesulfonyl chloride, phycoerythrin, lanthanide phosphors and the like. An antibody may also be labelled with enzymes that are useful for detection, such as horseradish peroxidase, galactosidase, luciferase, alkaline phosphatase, glucose oxidase and the like. When an antibody is labelled with a detectable enzyme, it is detected by adding additional reagents that the enzyme uses to produce a reaction product that can be discerned. For example, when the agent horseradish peroxidase is present, the addition of hydrogen peroxide and diaminobenzidine leads to a colored reaction product, which is detectable.
An antibody may also be labelled with biotin, and detected through indirect measurement

- 81 -of avidin or streptavidin binding. An antibody may be labelled with a magnetic agent, such as gadolinium. An antibody may also be labelled with a predetermined polypeptide epitope recognized by a secondary reporter (e. g. leucine zipper pair sequences, binding sites for secondary antibodies, metal binding domains, epitope tags). In some embodiments, labels may be attached by spacer arms of various lengths to reduce potential steric hindrance.
The antibody may also be derivatized with a chemical group such as polyethylene glycol (PEG), a methyl or ethyl group, or a carbohydrate group.
These groups may be useful to improve the biological characteristics of the antibody, e.g. to potentially increase serum half-life or to potentially increase tissue binding.
Antibody Specificity In some embodiments comprising antigen binding domains, at least the antigen binding domain (for example, but not limited to, an antibody variable region having all 6 CDRs, or an equivalent region that is at least 90 percent identical to an antibody variable region) may be chosen from that found in: abagovomab, abatacept (ORENCIA ), abciximab (REOPRO , c7E3 Fab), adalimumab (HUMIRAO), adecatumumab, alemtuzumab (CAMPATHO, MabCampath or Cannpath-1H), altumomab, afelimomab, anatumomab mafenatox, anetumumab, anrukizumab, apolizumab, arcitumomab, aselizumab, atlizumab, atorolimumab, bapineuzumab, basiliximab (SIMULECTe), bavituximab, bectumomab (LYMPHOSCAN ), belimumab (LYMPHO-STAT-B0), bertilimumab, besilesomab, I3cept (ENBRELO), bevacizumab (AVASTIN ), biciromab brallobarbital, bivatuzumab mertansine, brentuximab vedotin (ADCETRIS ), canakinumab (ACZ885), cantuzumab mertansine, capromab (PROSTASCINTO), catumaxomab (REMOV AB ), cedelizumab (CIMZIA0), certolizumab pegol, cetuximab (ERBITUX0), clenoliximab, dacetuzumab, dacliximab, daclizumab (ZENAP AX( ), denosumab (AMG 162), detumomab, dorlimomab aritox, dorlixizumab, duntumumab, durimulumab, durmulumab, ecromeximab, eculizumab (SOLIRIS ), edobacomab, edrecolomab (Mab17-1A, PANOREXO), efalizumab (RAPTIVAe), efungumab (MYCOGRAB ), elsilimomab, enlimomab pegol, epitumomab cituxetan, efalizumab, epitumomab, , ,

- 82 -epratuzumab, erlizumab, ertumaxomab (REXOMUNO), etaracizumab (etaratuzumab, VITAXIN , ABEGRINTm), exbivirumab, fanolesomab (NEUTROSPECO), faralimomab, felvizumab, fontolizumab (HUZAF ), galiximab, gantenerumab, gavilimomab (ABX-CBL(R)), gemtuzumab ozogamicin (MYLOTARGO), golimumab (CNTO 148), gomiliximab, ibalizumab (TNX-355), ibritumomab tiuxetan (ZEVALINO), igovomab, imciromab, infliximab (REMICAD E ), inolimonnab, inotuzumab ozogamicin, ipilimumab (YERVOY , MDX-010), iratumumab, keliximab, labetuzumab, lemalesomab, lebrilizumab, lerdelimumab, lexatumumab (HGS-ETR2, ETR2-ST01), lexitumumab, libivirumab, lintuzumab, lucatumumab, lumiliximab, mapatumumab (HGS-ETRI, TRM-I), maslimomab, matuzumab (EMD72000), mepolizumab (BOSATRIA0), metelimumab, milatuzunnab, minretumomab, mitumomab, morolimumab, motavizumab (NUMAXTm), muromonab (OKT3), nacolomab tafenatox, naptumomab estafenatox, natalizumab (TYSABRI , ANTEGRENO), nebacumab, nerelimomab, nimotuzumab (THERACIM hR30, THERA-CIM-hR30, THERALOCO), nofetumomab merpentan (VERLUMA0), ocrelizumab, odulimomab, ofatumumab, omalizumab (XOLAIRO), oregovomab (OVAREXO), otelixizumab, pagibaximab, palivizumab (SYNAGIS ), panitumumab (ABX-EGF, VECTIBIXO), pascolizumab, pemtumomab (THERAGYNO), pertuzumab (2C4, OMNITARGO), pexelizunnab, pintumomab, ponezumab, priliximab, pritumumab, ranibizumab (LUCENTIS ), raxibacumab, regavirumab, reslizumab, rituximab (RITUXAN , MabTHERAO), rovelizumab, ruplizumab, satumomab, sevirumab, sibrotuzumab, siplizumab (MEDI-507), sontuzumab, stamulumab (Myo-029), sulesomab (LEUKOSCANO), tacatuzumab tetraxetan, tadocizumab, talizumab, taplitumomab paptox, tefibazumab (AUREXIS ), telimomab aritox, teneliximab, teplizumab, ticilimumab, tocilizumab (ACTEMRA ), toralizumab, tositumomab, trastuzumab (HERCEPTINO), tremelimumab (CP-675,206), tucotuzumab celmoleukin, tuvirumab, urtoxazumab, ustekinumab (CNTO 1275), vapaliximab, veltuzumab, vepalimomab, visilizumab (NUVIONO), volociximab (M200), votumumab (HUMASPECTO), zalutumumab, zanolimumab (HuMAX-CD4), ziralimumab, or zolimomab aritox.

- 83 -In some embodiments comprising antigen binding domains, the antigen binding domain comprises a heavy and light chain variable domain having six CDRs, and/or may compete for binding with an antibody selected from the preceding list. In some embodiments comprising antigen binding domains, the antigen binding domain may bind to the same epitope as the antibodies in the preceding list. In some embodiments comprising antigen binding domains, the antigen binding domain comprises a heavy and light chain variable domain having six total CDRs, and may bind to the same antigen as the antibodies in the preceding list.
In some embodiments comprising antigen binding domains, at least the first antigen binding domain comprises a heavy and light chain variable domain having six (6) total CDRs, and may specifically bind to an antigen selected from: PDGFRa, PDGFRp, PDGF, VEGF, VEGF-A, VEGF-B, VEGF-C, VEGF-D, VEGF-E, VEGF-F, VEGFR1, VEGFR2, VEGFR3, FGF, FGF2, HGF, KDR, fit-1, FLK-1, Ang-2, Ang-1, PLGF, CEA, CXCL13, Baff, IL-21, CCL21, TNF-a, CXCL12, SDF-I, bFGF, MAC-I, IL23p19, FPR, IGFBP4, CXCR3, TLR4, CXCR2, EphA2, EphA4, EphrinB2, EGFR(ErbBI), HER2(ErbB2 or p185neu), HER3(ErbB3), HER4 ErbB4 or tyro2), SCI, LRP5, LRP6, RAGE, s100A8, s100A9, Nav1.7, GLPI, RSV, RSV F protein, Influenza HA protein, Influenza NA protein, HMGBI, CD16, CD19, CD20, CD21, CD28, CD32, CD32b, CD64, CD79, CD22, ICAM-1, FGFRI, FGFR2, HDGF, EphB4, GITR, 13 -amyloid, hMPV, PIV-I, PIV-2, OX4OL, IGFBP3, cMet, PD-I, PLGF, Neprolysin, CTD, IL- 18, IL-6, CXCL- 13, IL-IRI, IL-15, IL-4R, IgE, PAI-I, NGF, EphA2, uPARt, DLL-4, av135, av136, a5131, a3131, interferon receptor type land type 11, CD 19, ICOS, IL- 17, Factor 11, Hsp90, IGF, IGF-I, IGF-II, CD 19, GM-CSFR, PIV-3, CMV, IL- 13, IL-9, and EBV.
In some embodiments comprising antigen binding domains, at least the first antigen binding domain may specifically bind to a member (receptor or ligand) of the TNF superfamily. Various molecules may include, but are not limited to Tumor Necrosis Factor-a ("TNF-a"), Tumor Necrosis Factor-13 ("TNF-13"), Lymphotoxin-a ("LT-a"), CD30 ligand, CD27 ligand, CD40 ligand, 4-1 BB ligand, Apo-1 ligand (also referred to as Fas ligand or CD95 ligand), Apo-2 ligand (also referred to as TRAIL),

- 84 -Apo-3 ligand (also referred to as TWEAK), osteoprotegerin (OPG), APRIL, RANK
ligand (also referred to as TRANCE), TALL-I (also referred to as BlyS, BAFF or THANK), DR4, DR5 (also known as Apo-2, TRAIL-R2, TR6, Tango-63, hAP08, TRICK2, or KILLER), DR6, DcRI, DcR2, DcR3 (also known as TR6 or M68), CARI, HVEM (also known as ATAR or TR2), GITR, ZTNFR-5, NTR-I, TNFLI, CD30, LTBr, 4-1BB receptor and TR9.
In some embodiments comprising antigen binding domains, at least the first antigen binding domain may be capable of binding one or more targets chosen from 5T4, ABL, ABCB5, ABCFI, ACVRI, ACVRIB, ACVR2, ACVR2B, ACVRLI, AD0RA2A, Aggrecan, AGR2, AICDA, AIFI, AIGI, AKAPI, AKAP2, AMH, AMHR2, angiogenin (ANG), ANGPTI, ANGPT2, ANGPTL3, ANGPTL4, Annexin A2, ANPEP, APC, APOCI, AR, aromatase, ATX, AXI, AZGPI (zinc-a-glycoprotein), B7.1, B7.2, B7-H1, BAD, BAFF, BAGI, BAII, BCR, BCL2, BCL6, BDNF, BLNK, BLRI (MDR15), BlyS, BMP1, BMP2, BMP3B (GDF10), BMP4, BMP6, BMP7, BMP8, BMP9, BMP11, BMP12, BMPR1A, BMPR1B, BMPR2, BPAGI (plectin), BRCAI, C19orf10 (IL27w), C3, C4A, C5, C5R1, CANTI, CASPI, CASP4, CAVI, CCBP2 (D6 / JAB61), CCLI (1-309), CCLI 1 (eotaxin), CCL13 (MCP-4), CCL15 (MIP-Id), CCL16 (HCC-4), CCL17 (TARC), CCL18 (PARC), CCL19 (MIP-3b), CCL2 (MCP -1), MCAF, CCL20 (MIP-3a), CCL21 (MEP-2), SLC, exodus-2, CCL22(MDC / STC-I), CCL23 (MPIF- I), CCL24 (MPIF-2 /
eotaxin-2), CCL25 (TECK), CCL26(eotaxin-3), CCL27 (CTACK / ILC), CCL28, CCL3 (MIP-la), CCL4 (MIP-lb), CCL5(RANTES), CCL7 (MCP-3), CCL8 (mcp-2), CCNAI, CCNA2, CCNDI, CCNEI, CCNE2, CCRI (CKRI / HM145), CCR2 (mcp- IRB /
RA),CCR3 (CKR3 / CMKBR3), CCR4, CCR5(CMKBR5 / ChemR13), CCR6 (CMKBR6 / CKR-L3 / STRL22 / DRY6), CCR7 (CKR7 / EBI1),CCR8 (CMKBR8 /
TERI / CKR-LI), CCR9 (GPR-9-6), CCRLI (VSHKI), CCRL2 (L-CCR),CD164, CD19, CDIC, CD20, CD200, CD-22, CD24, CD28, CD3, CD33, CD35, CD37, CD38, CD3E, CD3G,CD3Z, CD4, CD40, CD4OL, CD44, CD45RB, CD46, CD52, CD69, CD72, CD74, CD79A, CD79B, CD8, CD80, CD81, CD83, CD86, CD105, CD137, CDHI (E-cadherin), CDCP1CDH10, CDH12, CDH13, CDH18,CDH19, CDH20, CDH5, CDH7, CDH8, CDH9, CDK2, CDK3, CDK4, CDK5, CDK6, CDK7, CDK9, CDKNIA

=

- 85 -(p21Wapl/Cipl), CDKNIB (p27Kipl), CDKNIC, CDKN2A (p16INK4a),CDKN2B, CDKN2C, CDKN3, CEBPB, CERI, CHGA, CHGB, Chitinase, CHSTIO, CKLFSF2,CKLFSF3, CKLFSF4, CKLFSF5, CKLFSF6, CKLFSF7, CKLFSF8, CLDN3, CLDN7 (claudin- 7), CLN3, CLU (clusterin), CMKLRI, CMKORI (RDCI), CNRI, COLI 8AI, COL1A1.COL4A3, COL6A1, CR2, Cripto, CRP, CSFI (M-CSF), CSF2 (GM-CSF), CSF3 (GCSF), CTLA4, CTL8, CTNNBI (b-catenin), CTSB
(cathepsin B), CX3CL1 (SCYDI), CX3CR1 (V28), CXCLI(GROD, CXCLIO (IP-10), CXCLII (I-TAC / IP-9), CXCL12 (SDFI), CXCL13, CXCL 14,CXCL 16, CXCL2 (GRO2), CXCL3 (GRO3), CXCL5 (ENA-78 / LIX), CXCL6 (GCP-2), CXCL9 (MIG), CXCR3 (GPR9/CKR-L2), CXCR4, CXCR6 (TYMSTR /STRL33 / Bonzo),CYB5, CYCI, Cyr61, CYSLTRI, c-Met, DAB2IP, DES, DKFZp451J0118, DNCLI, DPP4, E2F1, ECGFI5EDGI, EFNAI, EFNA3, EFNB2, EGF, ELAC2, ENG, endoglin, ENO', EN02, EN03, EPHAI, EPHA2, EPHA3, EPHA4, EPHA5, EPHA6, EPHA7, EPHA8, EPHA9, EPHAIO, EPHBI, EPHB2, EPHB3, EPHB4, EPHB5, EPHB6, EPHRIN-AI, EPHRIN-A2, EPHRIN-A3, EPHR1N-A4, EPHRIN-A5, EPHRIN-A6, EPHRIN-BI, EPHR1N-B2, EPHRTN-B3, EPHB4,EPG, ERBB2 (Her-2), EREG, ERK8, Estrogen receptor, ESRI, ESR2, F3 (IF), FADD, farnesyltransferase, FasL, FASNf, FCER1A,FCER2, FCGR3A, FGF, FGFI (aFGF), FGFIO, FGFI 1, FGF12, FGF12B, FGF13, FGF14, FGF16, FGF17, FGF18, FGF19, FGF2 (bFGF), FGF20, FGF21 (such as mimAb1), FGF22, FGF23, FGF3 (int-2),FGF4 (HST), FGF5, FGF6 (HST-2), FGF7 (KGF), FGF8, FGF9, FGFR3, FIGF (VEGFD), FILI(EPSILON), FBLI (ZETA), FLJ12584, FLJ25530, FLRTI (fibronectin), FLTI, FLT-3, FOS, FOSLI(FRA-1), FY (DARC), GABRP (GABAa), GAGEBI, GAGECI, GALNAC4S-6ST, GATA3, GD2, GD3, GDF5, GDF8, GFII, GGTI, GM-CSF, GNASI, GNRHI, GPR2 (CCRIO), GPR31, GPR44, GPR81 (FKSG80), GRCCIO (CIO), gremlin, GRP, GSN (Gelsolin), GSTPI, HAVCR2, HDAC, HDAC4, HDAC5,HDAC7A, HDAC9, Hedgehog, HGF, HIFIA, HIPI, histamine and histamine receptors, HLA-A, HLA-DRA, HM74, HMOXI, HSP90, HUMCYT2A, ICEBERG, ICOSL, ID2, IFN-a, IFNAI, IFNA2, IFNA4,IFNA5, EFNA6, BFNA7, IFNBI, IFNgamma, IFNWI, IGBPI, IGFI, IGFIR, IGF2, IGFBP2,IGFBP3, IGFBP6, DL-I, ILIO, 1LIORA, ILIORB, IL- 1, ILIRI (CD121a), ILIR2(CD121b), IL- IRA, IL-2, IL2RA
(CD25),

- 86 -IL2RB(CD122), IL2RG(CD132), IL-4, IL-4R(CD123), IL-5, IL5RA(CD125), IL3RB(CD131), IL-6, IL6RA (CD126), IR6RB(CD130), IL-7, IL7RA(CD127), IL-8, CXCRI (IL8RA), CXCR2 (IL8RB/CD128),IL-9, IL9R (CD129), IL- 10, 1L1ORA(CD210), IL1ORB(CDW210B), IL-11, ILI IRA, IL-12, 1L-12A, 1L-12B, IL-12RB1,1L-12RB2, IL-13, IL13RA1, IL13RA2, IL14, IL15, IL15RA, 1L16, IL17,1L17A, IL17B, IL17C, IL17R, IL18, IL18BP, IL18R1, IL18RAP, IL19, ILIA, ILIB, ILIFIO, !Li F5, IL1F6, IL1F7, IL1F8, DL1F9, ILIHYI, ILIRI, IL1R2, ILIRAP, ILIRAPLI, IL1RAPL2, ILIRLI, IL1RL2, ILIRN, IL2, IL20, IL2ORA, IL21R, IL22, IL22R, IL22RA2, IL23,DL24, IL25, IL26, IL27, IL28A, IL28B, IL29, IL2RA, IL2RB,IL2RG, IL3, IL30, IL3RA, IL4,1L4R,IL6ST (glycoprotein 130), ILK, INHA, INHBA, INSL3, INSL4, IRAKI, IRAK2, ITGA1, ITGA2, ITGA3, ITGA6 (a 6 integrin), ITGAV, ITGB3, ITGB4 (13 4 integrin), JAKI, JAK3, JTB, JUN,K6HF, KAII, KDR, KIM-1, KITLG, KLF5 (GC Box BP), KLF6, KLKIO, KLK12, KLK13, KLK14, KLK15, KLK3, KLK4, KLK5, KLK6, KLK9, KRTI, KRT19 (Keratin 19), KRT2A, KRTHB6 (hair-specific type II keratin), LAMA5, LEP
(leptin), Lingo- p75, Lingo-Troy, LPS, LRP5, LRP6, LTA (TNF- b), LTB, LTB4R
(GPR16), LTB4R2, LTBR, MACMARCKS, MAG or Omgp, MAP2K7 (c-Jun), MCP-I, MDK, MIBI, midkine, MIF, MISRII, MJP-2,MK, MKI67 (Ki-67), MMP2, MMP9, MS4A1, MSMB,MT3 (metallothionectin-Ui), mTOR, MTSSI, MUCI (mucin), MYC, MYD88, NCK2, neurocan, neuregulin-1, neuropilin-1, NFKBI, NFKB2, NGFB (NGF), NGFR, NgR-Lingo, NgR-Nogo66 (Nogo), NgR- p75, NgR-Troy, NMEI (NM23A), NOTCH, NOTCH!, NOX5, NPPB, NROBI, NROB2, NRIDI, NR1D2, NR1H2, NR1H3, NR1H4, NR1I2, NR1I3, NR2C1, NR2C2, NR2E1, NR2E3, NR2F1, NR2F2, NR2F6, NR3C1, NR3C2, NR4A1, NR4A2, NR4A3, NR5A1, NR5A2, NR6A1, NRPI, NRP2, NT5E, NTN4, OCT-1, ODZ1, OPN1, OPN2, OPRDI, P2RX7, PAP, PARTI, PATE, PAWR, PCA3, PCDGF, PCNA, PDGFA, PDGFB, PDGFRA, PDGFRB, PECAMI, peg-asparaginase, PF4 (CXCL4), Plexin B2 (PLXNB2), PGF, PGR, phosphacan, PIAS2, PI3 Kinase, PIK3CG, PLAU (uPA), PLG5PLXDCI, PKC, PKC-p, PPBP (CXCL7), PPID, PRI, PRKCQ, PRKDI, PRL, PROC, PROK2, pro-NGF, prosaposin, PSAP, PSCA, PTAFR, PTEN, PTGS2 (COX-2), PTN, RAC2 (P21Rac2), RANK, RANK
ligand, RARB, RGSI, RGS13, RGS3,RNFI10 (ZNF144), Ron, ROB02, RXR, selectin,

- 87 -S100A2, S100A8, S100A9, SCGB 1D2 (lipophilin B), SCGB2A1 (mammaglobin 2),SCGB2A2 (mammaglobin 1), SCYEI (endothelial Monocyte-activating cytokine), SDF2, SERPENA1, SERPINA3, SERPINB5 (maspin), SERPINEI (PAI-I), SERPINFI, SHIP-I, SHIP-2, SHBI, SHB2, SHBG, SfcAZ,SLC2A2, SLC33A1, SLC43A1, SLIT2, SPPI, SPRRIB (SprI), ST6GAL1, STAB!, STAT6, STEAP, STEAP2, SULF-1, SuIf-2, TB4R2, TBX21, TCPIO, TDGFI, TEK, TGFA, TGFBI, TGFBIII, TGFB2,TGFB3, TGFBI, TGFBRI, TGFBR2, TGFBR3, THIL, THBSI (thrombospondin-1), THBS2/THBS4, THPO, TIE (Tie-1), TIMP3, tissue factor, TIKI2, TLR10, TLR2, TLR3, TLR4, TLR5, TLR6JLR7, TLR8, TLR9, TM4SF1, TNF, TNF-a, TNFAIP2 (B94), TNFAIP3, TNFRSFIIA, TNFRSFIA, TNFRSFIB, TNFRSF21, TNFRSF5, TNFRSF6 (Fas), TNFRSF7, TNFRSF8, TNFRSF9, TNFSFIO (TRAIL), TNFSFI 1 (TRANCE), TNFSF12 (APO3L), TNFSF13 (April), TNFSF13B,TNFSF14 (HVEM-L), TNFSF15 (VEGI), TNFSF 18, TNFSF4 (0X40 ligand), TNFSF5 (CD40 ligand), TNFSF6 (FasL), TNFSF7 (CD27 ligand), TNFSF8 (CD30 ligand), TNFSF9 (4-1BB ligand), TOLLIP, Toll-like receptors, TLR2, TLR4, TLR9, TOP2A (topoisomerase lia), TP53, TPMI, TPM2,TRADD, TRAFI, TRAF2, TRAF3, TRAF4, TRAF5, TRAF6, TRKA, TREMI, TREM2, TRPC6, TROY, TSLP, TWEAK, Tyrosinase, uPAR, VEGF, VEGFB, VEGFC, versican, VHL C5, VLA-4, Wnt-1, XCLI (Iymphotactin), XCL2 (SCM-Ib), XCRI (GPR5 /
CCXCRI), YYI, and ZFPM2.
DEFINITIONS
Generally, nomenclatures used in connection with, and techniques of, biochemistry, analytical chemistry, synthetic organic chemistry, medicinal and pharmaceutical chemistry, cell and tissue culture, molecular biology, immunology, microbiology, genetics and protein and nucleic acid chemistry and hybridization described herein are those well known and commonly used in the art. The methods and techniques of the present invention are generally performed according to conventional methods well known in the art and as described in various general and more specific references that are cited and discussed throughout the present specification unless otherwise indicated. Reactions and purification techniques are performed according to manufacturer's specifications, as commonly accomplished in

- 88 -the art or as described herein. As used herein, the 20 natural, or conventional, amino acids and their abbreviations follow IUPAC single letter and three letter codes.
A "complementary residue set", as used herein, refers to at least one amino acid in a CH-1 domain, and at least one amino acid in the CL domain that are engineered to interact with each other. By interacting with each other, they drive their respective domains to heterodimerize and form an interface comprising at least some of the interaction between the residues of the complementary residue set. The interaction may be characterized by a salt bridge, electrostatic interaction, or van der Waals force. A complementary residue set may comprise more than one engineered residue in each domain.
Any given residue within a complementary residue set will be within 5A of at least one other residue of that complementary residue set.
In the context of complementary residue sets, two residues are said to interact if at least one atom of each residue is within 5A of each other. Residue interaction may be characterized as either a salt bridge, electrostatic interaction, or van der Waals force. For avoidance of doubt, in other contexts it is recognized that interatomic forces may act over longer distances.
"Complementary pairing" between domains refers to the interaction of those two domains, at least in part, through a complementary residue set.
"Engineered", as used herein, refers to the deliberate mutation of residues that are not found in the predominant wild type sequence, and may be an engineered insertion, deletion or substitution mutation.
A "heteromultimer", "heteromultimeric complex", or "heteromultimeric polypeptide" is a molecule comprising at least a first polypeptide and a second polypeptide, wherein the second polypeptide differs in amino acid sequence from the first polypeptide by at least one amino acid residue. The heteromultimer can comprise a "heterodimer" formed by the first and second polypeptide or can form higher order tertiary structures where polypeptides in addition to the first and second polypeptide are present.
A "heterodimer," "heterodimeric protein," "heterodimeric complex," or

- 89 -"heteromultimeric polypeptide" is a molecule comprising a first polypeptide and a second polypeptide, wherein the second polypeptide differs in amino acid sequence from the first polypeptide by at least one amino acid residue.
In the context of the invention, the term heterodimeric is used to indicate a heteromultimer comprising at least two polypeptides with differing amino acid sequences; but it will be readily appreciated that in many embodiments, particularly those where the invention relates to IgG antibodies and similar molecules, heterodimeric proteins of the invention may equally be referred to as heteromultimeric proteins, as there will necessarily be four distinct polypeptides (the first heavy and light chain, and the second heavy and light chain).
"Polypeptide," "peptide," and "protein" are used interchangeably to refer to a polymer of amino acid residues. As used herein, these terms apply to amino acid polymers in which one or more amino acid residues is an artificial chemical analog of a corresponding naturally occurring amino acid. These terms also apply to naturally occurring amino acid polymers. Amino acids can be in the L-form or D-form as long as the binding and other desired characteristics of the peptide are maintained. A
polypeptide may be monomeric or polymeric. The terms also encompass an amino acid chain that has been modified naturally or by intervention; for example, disulfide bond formation, glycosylation, lipidation, acetylation, phosphorylation, or any other manipulation or modification, such as conjugation with a labeling component.
Also included within the definition are, for example, polypeptides containing one or more analogs of an amino acid (including, for example, unnatural amino acids, etc.), as well as other modifications known in the art. It is understood that the polypeptides can occur as single chains or associated chains.
Unless indicated otherwise by a "D" prefix, e.g. D-Ala or N-Me-D-11e, or written in lower case format, e.g. a, i, I, (D versions of Ala, Ile, Leu), the stereochemistry of the a-carbon of the amino acids and aminoacyl residues in peptides described in this specification and the appended claims is the natural or "L" configuration.
All peptide sequences are written according to the generally accepted convention whereby the a-N-terminal amino acid residue is on the left and the a-C-

- 90 -terminal amino acid residue is on the right. As used herein, the term "N-terminus"
refers to the free a-amino group of an amino acid in a peptide, and the term "C-terminus" refers to the free a-carboxylic acid terminus of an amino acid in a peptide.
A peptide which is N-terminated with a group refers to a peptide bearing a group on the a-amino nitrogen of the N-terminal amino acid residue. An amino acid which is C-terminated with a group refers to an amino acid bearing a group on the carboxyl moiety, such as a methyl group resulting in a methyl ester.
As used herein, "biological activity" refers to the potential in vivo activities of a compound, composition, or other mixture, or physiological responses that may result upon in vivo administration of a compound, composition or other mixture.
The term "biologically compatible" as used herein means something that is biologically inert or non reactive with intracellular and extra cellular biological molecules, and non toxic.
"About" or "approximately," when used in connection with a measurable numerical variable, refers to the indicated value of the variable and to all values of the variable that are within the experimental error of the indicated value (e.g.
within the 95% confidence interval for the mean) or within 10 percent of the indicated value, whichever is greater. Numeric ranges are inclusive of the numbers defining the range.
The term "identity," as known in the art, refers to a relationship between the sequences of two or more polypeptide molecules or two or more nucleic acid molecules, as determined by comparing the sequences. In the art, "identity"
also means the degree of sequence relatedness between polypeptide or nucleic acid molecule sequences, as the case may be, as determined by the match between strings of nucleotide or amino acid sequences."Identity" measures the percent of identical matches between two or more sequences with gap alignments addressed by a particular mathematical model of computer programs (i. e. "algorithms").
The term "similarity" is a related concept, but in contrast to "identity", refers to a measure of similarity which includes both identical matches and conservative substitution matches. Since conservative substitutions apply to polypeptides and not ,

- 91 -nucleic acid molecules, similarity only deals with polypeptide sequence comparisons.
If two polypeptide sequences have, for example, 10 out of 20 identical amino acids, and the remainder are all nonconservative substitutions, then the percent identity and similarity would both be 50%. If in the same example, there are 5 more positions where there are conservative substitutions, then the percent identity remains 50%, but the percent similarity would be 75% (15 out of 20). Therefore, in cases where there are conservative substitutions, the degree of similarity between two polypeptide sequences will be higher than the percent identity between those two sequences.
The term "conservative amino acid substitution" refers to a substitution of a native amino acid residue with a nonnative residue such that there is little or no effect on the polarity, charge, and approximate volume of the amino acid residue at that position. For example, a conservative substitution results from the replacement of a non-polar residue in a polypeptide with any other non-polar residue. The term may also refer to a substitution identified as frequently occurring between highly similar proteins, as in the BLOSUM62 matrix or related matrices (Proc. Natl. Acad.
Sci. USA
89(22), 10915-9, 1992).
The term "vector," as used herein, is intended to refer to a nucleic acid molecule capable of transporting another nucleic acid to which it has been linked.
One type of vector is a "plasmid," which refers to a circular double stranded DNA loop into which additional DNA segments may be ligated. Another type of vector is a phage vector. Another type of vector is a viral vector, wherein additional DNA

segments may be ligated into the viral genome. Certain vectors are capable of autonomous replication in a host cell into which they are introduced (e.g.
bacterial vectors having a bacterial origin of replication and episomal mammalian vectors).
Other vectors (e.g. non-episomal mammalian vectors) can be integrated into the genome of a host cell upon introduction into the host cell, and thereby are replicated along with the host genome. Moreover, certain vectors are capable of directing the expression of genes to which they are operatively linked. Such vectors are referred to herein as "recombinant expression vectors" (or simply, "recombinant vectors").
In general, expression vectors of utility in recombinant DNA techniques are often in the

- 92 -form of plasmids. In the present specification, "plasmid" and "vector" may be used interchangeably as the plasmid is the most commonly used form of vector.
"Polynucleotide," or "nucleic acid molecule," which may be used interchangeably herein, refers to a polymeric, possibly isolated, form of nucleosides or nucleotides of at least 10 bases in length. The term includes single and double stranded forms. The nucleotides can be deoxyribonucleotides, ribonucleotides, modified nucleotides or bases, and/or their analogs, or any substrate that can be incorporated into a polymer by DNA or RNA polymerase, or by a synthetic reaction.
A polynucleotide may comprise modified nucleotides, such as methylated nucleotides and their analogs. If present, modification to the nucleotide structure may be imparted before or after assembly of the polymer. The sequence of nucleotides may be interrupted by non-nucleotide components. A polynucleotide may be further modified after synthesis, such as by conjugation with a label. Other types of modifications include, for example, "caps", substitution of one or more of the naturally occurring nucleotides with an analog, internucleotide modifications such as, for example, those with uncharged linkages (e.g. methyl phosphonates, phosphotriesters, phosphoamidates, carbamates, etc.) and with charged linkages (e.g. phosphorothioates, phosphorodithioates, etc.), those containing pendant moieties, such as, for example, proteins (e.g. nucleases, toxins, antibodies, signal peptides, ply-L-lysine, etc.), those with intercalators (e.g. acridine, psoralen, etc.), those containing chelators (e.g. metals, radioactive metals, boron, oxidative metals, etc.), those containing alkylators, those with modified linkages (e.g. alpha anomeric nucleic acids, etc.), as well as unmodified forms of the polynucleotide(s).
Further, any of the hydroxyl groups ordinarily present in the sugars may be replaced, for example, by phosphonate groups, phosphate groups, protected by standard protecting groups, or activated to prepare additional linkages to additional nucleotides, or may be conjugated to solid or semi-solid supports. The 5' and 3' terminal OH can be phosphorylated or substituted with amines or organic capping group moieties of from 1 to 20 carbon atoms. Other hydroxyls may also be derivatized to standard protecting groups. Polynucleotides can also contain analogous forms of ribose or deoxyribose

- 93 -sugars that are generally known in the art, including, for example, 2'-0-methyl-, 2'-0-allyl, 2'-fluoro- or 2'-azido-ribose, carbocyclic sugar analogs, alpha-anomeric sugars, epimeric sugars such as arabinose, xyloses or lyxoses, pyranose sugars, furanose sugars, sedoheptuloses, acyclic analogs and abasic nucleoside analogs such as methyl riboside. One or more phosphodiester linkages may be replaced by alternative linking groups. These alternative linking groups include, but are not limited to, embodiments wherein phosphate is replaced by P(0)S("thioate"), P(S)S
("dithioate"), "(0)NR2 ("amidate"), P(0)R, P(0)OR', CO or CH2 ("formacetal"), in which each R
or R' is independently H or substituted or unsubstituted alkyl (1-20 C.) optionally containing an ether (--0--) linkage, aryl, alkenyl, cycloalkyl, cycloalkenyl or araldyl.
Not all linkages in a polynucleotide need be identical. The preceding description applies to all polynucleotides referred to herein, including RNA and DNA.
"Oligonucleotide," as used herein, generally refers to short, generally single stranded, generally synthetic polynucleotides that are generally, but not necessarily, less than about 200 nucleotides in length. The terms "oligonucleotide" and "polynucleotide" are not mutually exclusive. The description above for polynucleotides is equally and fully applicable to oligonucleotides.
A reference to a nucleotide sequence as used herein encompasses its complement unless otherwise specified. Thus, a reference to a nucleic acid having a particular sequence should be understood to encompass its complementary strand, with its complementary sequence, unless otherwise defined by context.
"Cell" or "cell line," as used herein, includes various types of cells that can be used to express a heterodimeric protein, a polypeptide or a nucleic acid of the invention, e.g. prokaryotic cells, eukaryotic cells, mammalian cells, rat cells, human cells.
The term "purify," and grammatical variations thereof, is used to mean the removal, whether completely or partially, of at least one impurity from a mixture containing the polypeptide and one or more impurities, which thereby improves the level of purity of the polypeptide in the composition (i.e. by decreasing the amount (ppm) of impurity(ies) in the composition).

,

- 94 -The terms "ion-exchange" and "ion-exchange chromatography" refer to a chromatographic process in which an ionizable solute of interest (e.g. a protein of interest in a mixture) interacts with an oppositely charged ligand linked (e.g. by covalent attachment) to a solid phase ion exchange material under appropriate conditions of pH and conductivity, such that the solute of interest interacts non-specifically with the charged compound more or less than the solute impurities or contaminants in the mixture. The contaminating solutes in the mixture can be washed from a column of the ion exchange material or are bound to or excluded from the resin, faster or slower than the solute of interest. "Ion-exchange chromatography"
specifically includes cation exchange, anion exchange, and mixed mode chromatographies.
The term "immune effector cell" or "effector cell" as used herein refers to a cell within the natural repertoire of cells in the human immune system which may be activated to affect the viability of a target cell. The viability of a target cell may include cell survival, proliferation, and/or ability to interact with other cells.
Salt bridges are a type of noncovalent interaction. A salt bridge involves close-range direct interactions between two atoms with opposite formal charges. In the context of protein structure, salt bridges most often form between the anionic carboxylate (RC00-) of either aspartic acid or glutamic acid and the cationic ammonium (RNH3+) from lysine or the guanidinium (RNHC(NH2)2+) of arginine, with histidine another possibility. However, other amino acids may participate depending on changes to their pKa values and locations in the polypeptide chain (the N
and C
terminal residues may be ionized, and thus capable of salt bridge formation, regardless of amino acid type).
Electrostatic interactions are noncovalent interactions between atoms having nonzero charge. They may have favorable, unfavorable, or neutral interaction energies and may involve atoms which have formal charges, or which are polarized despite the lack of formal charge. Hydrogen bonds, salt bridges, and pi-cation stacking are examples of electrostatic interactions frequently observed in protein structures.

- 95 -Structural alignments, which are usually specific to protein and sometimes RNA sequences, use information about the secondary and tertiary structure of the protein or RNA molecule to aid in aligning the sequences. These methods are used for two or more sequences and typically produce local alignments; however, because they depend on the availability of structural information, they can only be used for sequences whose corresponding structures are known (usually through X-ray crystallography or NMR spectroscopy). Because both protein and RNA structure is more evolutionarily conserved than sequence, structural alignments can be more reliable between sequences that are very distantly related and that have diverged so extensively that sequence comparison cannot reliably detect their similarity.
Where there is no available structural data on one of the proteins, a comparison can still be made if structural data is available on one or preferably more closely related proteins, such as immunoglobulins across species, and in particular antibody constant domains across species and subtype.
Structural alignments are used as the "gold standard" because they explicitly align regions of the protein sequence that are structurally similar rather than relying exclusively on sequence information. A commonly used algorithm for structural alignments is TM-ALIGN (Zhang and Skolnick, Nucleic Acids Research, 33: 2302-2309 (2005)), which assigns increased weight to the most similar regions of the structure during superposition.
SEQUENCE ALIGNMENT
Where structural alignment with protein sequences of the invention is not possible, for example due to an absence of target sequence NMR or crystal structure data, sequence alignment may be used. The skilled person is familiar with sequence alignment tools (such as BLAST, CLUSTAL and others known to the skilled person, such as those described herein), and is able to align sequences, particularly antibody constant domain sequences according to known structural motifs, especially due to the large number of exemplary structural studies already existent for immunoglobulin domains, antibodies and antibody constant domains in particular, across subtype and species.

=

- 96 -Computational approaches to sequence alignment generally fall into two categories: global alignments and local alignments. Calculating a global alignment is a form of global optimization that "forces" the alignment to span the entire length of all query sequences. By contrast, local alignments identify regions of similarity within long sequences that are often widely divergent overall. Local alignments are often preferable, but can be more difficult to calculate because of the additional challenge of identifying the regions of similarity. A variety of computational algorithms have been applied to the sequence alignment problem. These include slow but formally correct methods like dynamic programming and also efficient, heuristic algorithms or probabilistic methods designed for large-scale database search, that do not guarantee to find best matches.
Global alignments, which attempt to align every residue in every sequence, are most useful when the sequences in the query set are similar and of roughly equal size. A general global alignment technique is the Needleman-Wunsch algorithm, which is based on dynamic programming. Local alignments are more useful for dissimilar sequences that are suspected to contain regions of similarity or similar sequence motifs within their larger sequence context. The Smith-Waterman algorithm is a general local alignment method also based on dynamic programming.
Pairwise sequence alignment methods are used to find the best-matching piecewise (local) or global alignments of two query sequences. The three primary methods of producing pairwise alignments are dot-matrix methods, dynamic programming, and word methods; however, multiple sequence alignment techniques can also align pairs of sequences. Although each method has its individual strengths and weaknesses, all three pairwise methods have difficulty with highly repetitive sequences of low information content - especially where the number of repetitions differ in the two sequences to be aligned. One way of quantifying the utility of a given pairwise alignment is the 'maximum unique match' (MUM), or the longest subsequence that occurs in both query sequences. Longer MUM sequences typically reflect closer relatedness. Preferred methods to determine identity and/or similarity are designed to give the largest match between the sequences tested. Methods to

- 97 -determine identity and similarity are codified in publicly available computer programs.
Preferred computer program methods to determine identity and similarity between two sequences include, but are not limited to, the GCG program package, including GAP (Devereux et al., Nuc. Acids Res. 12: 387 (1984); Genetics Computer Group, University of Wisconsin, Madison, WI), BLASTP, BLASTN, and FASTA (Atschul et al., J. Mol. Biol. 215: 403-10 (1990)). The BLAST X program is publicly available from the National Center for Biotechnology Information (NCBI) and other sources (Altschul et al., BLAST Manual (NCB NLM NIH, Bethesda, MD); Altschul et al., 1990, supra).
The well-known Smith Waterman algorithm may also be used to determine identity.
By way of example, using the computer algorithm GAP (Genetics Computer Group), two polypeptides for which the percent sequence identity is to be determined are aligned for optimal matching of their respective amino acids (the "matched span", as determined by the algorithm). A gap opening penalty (which is calculated as the average diagonal; the "average diagonal" is the average of the diagonal of the comparison matrix being used; the "diagonal" is the score or number assigned to each perfect amino acid match by the particular comparison matrix) and a gap extension penalty (which is usually 0.1X the gap opening penalty), as well as a comparison matrix such as PAM 250 or BLOSUM 62 are used in conjunction with the algorithm. Preferred parameters for polypeptide sequence comparison include the following: Algorithm: Needleman and Wunsch, J. Mol. Biol. 48: 443-53 (1970).
Comparison matrix: BLOSUM 62 from Henikoff et al., Proc. Natl. Acad. Sci. U.
S. A.
89: 10915-19 (1992).
Other exemplary algorithms, gap opening penalties, gap extension penalties, comparison matrices, thresholds of similarity, etc. may be used by those of skill in the art, including those set forth in the Program Manual, WisconsinPackage, Version 9, September, 1997. The particular choices to be made will depend on the specific comparison to be made, such as DNA to DNA, protein to protein, protein to DNA;
and additionally, whether the comparison is between given pairs of sequences (in which case GAP or BestFit are generally preferred) or between one sequence and a large database of sequences (in which case FASTA or BLASTA are preferred).

=

- 98 -For specific protein families with conserved structure, other alignment algorithms are available. In the case of antibodies, various algorithms for assigning Kabat numbering are available. The algorithm implemented in the 2012 release of Abysis (www.abysis.org) is used herein to assign Kabat numbering to variable regions unless otherwise noted.
The term "percent sequence identity" in the context of nucleic acid sequences means the residues in two sequences that are the same when aligned for maximum correspondence. The length of sequence identity comparison may be over a stretch of at least about nine nucleotides, usually at least about 18 nucleotides, more usually at least about 24 nucleotides, typically at least about 28 nucleotides, more typically at least about 32 nucleotides, and preferably at least about 36, 48 or more nucleotides.
There are a number of different algorithms known in the art which can be used to measure nucleotide sequence identity. For instance, polynucleotide sequences can be compared using FASTA, Gap or Bestfit, which are programs in Wisconsin Package Version10.0, Genetics Computer Group (GCG), Madison, Wisconsin.
FASTA, which includes, e.g. the programs FASTA2 and FASTA3, provides alignments and percent sequence identity of the regions of the best overlap between the query and search sequences (Pearson, Methods Enzymol. 183:63-98 (1990);
Pearson, Methods Mol. Biol. 132:185-219 (2000); Pearson, Methods Enzymol.
266:227-258 (1996); Pearson, J. MoL Biol. 276:71-84 (1998); incorporated herein by reference). Unless otherwise specified, default parameters for a particular program or algorithm are used. For instance, percent sequence identity between nucleic acid sequences can be determined using FASTA with its default parameters (a word size of 6 and the NOPAM factor for the scoring matrix) or using Gap with its default parameters as provided in GCG Version 6.1, incorporated herein by reference.
SEQUENCE LIST
SEQ Description Sequence 1 IgG1 CH1 numbered ---ASTKGPS VFPLAPSSKS --TSGGTAAL
GCLVKDYFPE PVTV SW
from residue 111 to N---SGALT SG-VHTFPAV LQS-SGLYSL SSVVTVPSSS LGT--Q-TYI
231 (with CNVNHKPSNT KVDKKV--EP KSC-insertions 162A, 162B, and 162C
after position 162) 2 IgG1 CH1 G1m3 using ---ASTKGPS VFPLAPSSKS --TSGGTAAL GCLVKDYFPE
PVTV-SW--- -*
%

- 99 -the numbering range N---SGALT SG-VHTFPAV LQS-SGLYSL SSVVTVPSSS LGT--Q-TYI
of Seq 1 CNVNHKPSNT KVDKRV--EP KSC-3 Ig52 CH1 using the ASTKGPS VFPLAPCSRS --TSESTAAL GCLVKDYFPE
PVTV-SW--- -numbering range of N---SGALT SG-VHTFPAV LQS-SGLYSL SSVVTVPSSN
FGT--Q-TYT
Seq 1 CNVDHKPSNT KVDKTV--ER K---4 IgG2 CH1 G2m23 ---ASTKGPS VFPLAPCSRS --TSESTAAL
GCLVKDYFPE PVTV-SW--- -using the numbering N---SGALT SG-VHTFPAV LQS-SGLYSL SSVVTVTSSN FGT--Q-TYT
range of Seq 1 CNVDHKPSNT KVDKTV--ER K---IgG3 CH1 using the ASTKGPS VFPLAPCSRS --TSGGTAAL GCLVKDYFPE PVTV-SW--- -numbering range of N---SGALT SG-VHTFPAV LQS-SGLYSL SSVVTVPSSS
LGT--Q-TYT
Seq 1 CNVNHKPSNT KVDKRV--EL KTP-6 IgG4 CH1 using the ASTKGPS VFPLAPCSRS --TSESTAAL GCLVKDYFPE
PVTV-SW--- -numbering range of N---SGALT SG-VHTFPAV LQS-SGLYSL SSVVTVPSSS
LGT--K-TYT
Seq 1 CNVDHKPSNT KVDKRV--ES KYG-7 IgM CH1 seql GSASAPT LFPLVSCENS P-SDTSSVAV GCLAQDFLPD
SITL-SW--- -numbered from KYKNNSDIS S--TRGFPSV LRG--GKYAA TSQVLLPSKD
VMQGTDEHVV
residue 111 to 226 CKVQH-PNGN --KEKNVPLP
(with insertions 162A, 162B, and 162C after position 162) 8 IgM CH1 seq2 using ---GSASAPT LFPLVSCENS P-SDTSSVAV
GCLAQDFLPD SITE-SW---- -the numbering range KYKNNSDIS S--TRGFPSV LRG--GKYAA TSQVLLPSKD VMAGTDEHVV
of Seq 7 CKVQH-PNGN --KEKNVPLP
9 CL-Kappa-KM3 RTV AAPSVFIFPP SDEQLKSGTA SVVCLLNNFY PREAKVQWKV
numbered from DNALQSGNSQ ESVTEQDSKD STYSLSSTLT
LSKADYEKHK VYACEVTHQG
position 101 to 215 LSSPVTKSFN RGEC-_ CL-Kappa-KM1 using RTV
AAPSVFIFPP SDEQLKSGTA SVVCLLNNFY PREAKVQWKV
the numbering range DNVLQSGNSQ ESVTEQDSKD STYSLSSTLT LSKADYEKHK LYACEVTHQG
of Seq 9 LSSPVTKSFN RGEC-11 CL-Kappa-KM1,2 RTV AAPSVFIFPP SDEQLKSGTA SVVCLLNNFY PREAKVQWKV
using the numbering DNALQSGNSQ ESVTEQDSKD STYSLSSTLT LSKADYEKHK LYACEVTHQG
range of Seq 9 LSSPVTKSFN RGEC-12 CL-Lambda using the QPK AAPSVTLFPP SSEELQANKA TLVCLISDFY PGAVTVAWKA
numbering range of DSSPVKAGVE TTTPSKQS-N NKYAASSYLS
LTPEQWKSHR SYSCQVTHEG --Seq 9 STVEKTVA PTECS
13 IgG1 CH2 numbered ---APELLGG PSVFLFPPKP KDTLMI-SRT
PEVTCVVVDV SHEDPEVKFN
from 241 to 360B WYV--DG--V EVR-NAKTKP REEQYN
STYRVVSVLT VLHQDWLNGK
with insertions EYKCKVSNKA LPAPIEKTI- SKAK--266A, 302A, 316A, 316B, 360A, and 14 IgG2 CH2 using the ---APPVA-G PSVFLFPPKP KDTLMI-SRT
PEVTCVVVDV SHEDPEVQFN
numbering range of WYV--DG--V EVH-NAKTKP REEQFN----STFRVVSVLT VVHQDWLNGK
Seq 13 EYKCKVSNKG LPAPIEKTI- SKTK--IgG3 CH2 using the ---APELLGG PSVFLFPPKP KDTLMI-SRT PEVTCVVVDV SHEDPEVQFK
numbering range of WYV--DG--V EVH-NAKTKP REEQYN----STFRVVSVLT VLHQDWLNGK
Seq 13 EYKCKVSNKA LPAPIEKTI- SKTK--16 IgG4 CH2 using the ---APEFLGG PSVFLFPPKP KDTLMI-SRT
PEVTCVVVDV SQEDPEVQFN
numbering range of WYV--DG--V EVH-NAKTKP REEQFN----STYRVVSVLT VLHQDWLNGK
Seq 13 EYKCKVSNKG LPSSIEKTI- SKAK--17 IgM CH2 using the VIAELP PKVSVFVPPR DGFFGN-PRK SKLICQATGF
S--PRQIQVS
numbering range of WLR--EG--K QVGSGVTTDQ VQAEAKESGP
TTYKVTSTLT IKESDWLGQS
Seq 13 MFTCRVDHRG L--TFQQNA- SSMCVP

=

- 100 -18 IgG1 CH3 numbered G-QPREPQVY TLPPSREE-- MTKNQVSLTC LVKGFYPS-D
IAV--EWES-from 361 to 478 NG--QPENNY KTTPPVLDS- D--GSFFLYS KLTVDKSRWQ
QGNVFSCSVM
with insertion 398A HEALHNHYTQ KSLSLSPGK
19 IgG1 C53 alternate G-QPREPQVY TLPPSRDE-- LTKNQVSLTC LVKGFYPS-D
IAV--EWES-isotype using the NG--QPENNY KTTPPVLDS- D--GSFFLYS KLTVDKSRWQ
QGNVFSCSVM
numbering range of HEALHNHYTQ KSLSLSPGK
Seq 18 20 IgG2 CH3 using the G-QPREPQVY TLPPSREE-- MTKNQVSLTC LVKGFYPS-D
IAV--EWES-numbering range of NG--QPENNY KTTPPMLDS- D--GSFFLYS KLTVDKSRWQ
QGNVFSCSVM
Seq 18 HEALHNHYTQ KSLSLSPGK
21 IgG3 CH3 using the G-QPREPQVY TLPPSREE-- MTKNQVSLTC LVKGFYPS-D
IAV--EWES-numbering range of SG--QPENNY NTTPPMLDS- D--GSFFLYS KLTVDKSRWQ
QGNIFSCSVM
Seq 18 HEALHNRFTQ KSLSLSPGK
22 IgG4 CH3 using the G-QPREPQVY TLPPSQEE-- MTKNQVSLTC LVKGFYPS-D
IAV--EWES-numbering range of NG--QPENNY KTTPPVLDS- D--GSFFLYS RLTVDKSRWQ
EGNVFSCSVM
Seq 18 HEALHNHYTQ KSLSLSLGK
23 IgM CH3 using the D-QDTAIRVF AIPPSFASI- FLTKSTKLTC LVTDLTTYDS
VTI--SWTRQ
numbering range of NG--EAV-KT HTNISESHP- N--ATFSAVG EASICEDDWN
SGERFTCTVT
Seq 18 HTDLPSP-LK QTISRPK--PREAKVQWKV
using the numbering DNALQSGNSQ ESVTEQDSKD STYSLDSTLT LSKADYEKHK VYACEVTHQG
range of Seq 9 LSSPVTKSFN RGEC

PREAKVQWKV
using the numbering DNALQSGNSQ ESVTEQDSKD STYSLRETLT LSKADYEKHK VYACEVTHQG
range of Seq 9 LSSPVTKSFN RGEC
26 CL-L1.1 using the RTV AAPSVFIFPP SDKQLKSGTA SVVCILNNFY
PREAKVQWKV
numbering range of DNALQSGNSQ ESVTEQDSKD STYSLCSTLT LSKADYEKHK
VYACEVTHQG
Seq 9 LSSPVTKSFN RGES
27 CL-L4.1 using the RTV AAPSVFIFPP SDCQLKSGTA HVVCLLNNFY
PREAKVQWKV
numbering range of DNALQSGNSQ ESVTEQDSKD STYSLSSTLT LSKADYEKHK
VYACEVTHQG
Seq 9 LSSPVTKSFN RGES
28 CL-14.2 using the RTV AAPSVFIFPP SDEQLKSGTA HVVCILNNFY
PREAKVQWKV
numbering range of DNALQSGNSQ ESVTEQDSKD STYSLCSTLT LSKADYEKHK
VYACEVTHQG
Seq 9 LSSPVTKSFN RGES
29 CL-L4.3 using the RTV AAPSVCIFPP SDEQLKSGTA HVVCLLNNFY
PREAKVQWKV
numbering range of DNALQSGNSQ ESVTEQDSKD STYSLSSTLT LSKADYEKHK
VYACEVTHQG
Seq 9 LSSPVTKSFN RGES
30 CL-H10.1 using the RTV AAPSVFIFPP SDEQLKSGTA DVSCLLNNFY
PREAKVQWKV
numbering range of DNALQSGNSQ ESVTEQDSKD STYSLCSSLT LSKADYEKHK
VYACEVTHQG
Seq 9 LSSPVTKSFN RGES
31 CL-H10.4 using the RTV AAPSVFIFPP SDCQLKSGTA SVMCLLNNFY
PREAKVQWKV
numbering range of DNALQSGNSQ ESVTEQDSKD STYSLGSGLT LSKADYEKHK
VYACEVTHQG
Seq 9 LSSPVTKSFN RGES
32 CL-5.6 using the RTV AAPSVCIFPP SDCQLKSGTA DVSCLLENFY
PREAKVQWKV
numbering range of DNALQSGNSQ ESVTEQDSKD STYSLCSSLT LSKADYEKHK
VYACEVTHQG
Seq 9 LSSPVTKSFN RGES

PVTV-SW--- -numbered from N---SGALT SG-VHTFPAV LQS-SGLYSL SSSVTVPSSS
LGT Q TYI
residue 111 to 230 CNVNHKPSNT KVDKKV--EP KSC
(with insertions 162A, 162B, and 162C after position 162) PVTV-SW--- -

- 101 -numbered using the N---SGALT SG-VHTFPAV LQS-SGLYSL GSVVTVPSSS LGT--Q-TYI
range of Seq 33 CNVNHKPSNT KVDKKV--EP KSC
35 CH1-L1.1 numbered ASTKGPS VFPLAPSSKS --TSGGTAAL GCLVKDYFPE PVTV-SW--- -using the range of N SGALT SG-VETCPAV LQS-SGLYSL SSIVTVPSSS LGT--Q-TYI
Seq 33 CNVNHKPSNT KVDDKV--EP KSS
36 CH1-L4.1 numbered ---ASTKGPS VCPLAPSSKS --TSGGTAAL GCLVEDYFPE PVTV-SW--- -using the range of N---SGALT SG-VHTFPAV LQS-SGLYSL SSVVTVPSSS LGT--Q-TYI
Seq 33 CNVNHKPSNT KVDKKV--EP KSS
37 CH1-L4.2 numbered ---ASTKGPS VFPLAPSSKS --TSGGTAAL GCHVKDYFPE PVTV-SW--- -using the range of N---SGALT SG-VHTCPAV LDS-SGLYEL SSIVTVPSSS LGT--Q-TYI
Seq 33 CNVNHKPSNT KVDKKV--EP KSS
38 CH1-L4.3 numbered ---ASTKGPS VFPLAPSSKS --TSGGTACL GCLVSDYFPE PVTV-SW--- -using the range of N---SGALT SG-VHTFPAV LQS-SGLYEL SSVVTVPSSS LGT--Q-TYI
Seq 33 CNVNHKPSNT KVDKKV--EP KSS
39 CH1-H10.1 numbered ---ASTKGPS VFPLAPSSKS --TSGGTAAL GCSVKDYFPE PVTV-SW--- -using the range of N---SGALT SG-VHTCPAV LQS-SGLYSL WSVVTVPSSS LGT--Q-TYI
Seq 33 CNVNHKPSNT KVDKKV--EP KSS
40 CH1-H10.4 numbered ---ASTKGPS VCPLAPSSKS --TSGGTAAL GCSVKDYFPE PVTV
SW
using the range of N SGALT SG-VHTFPAV LQS-SGLYSL WSVVTVPSSS LGT--Q-TYI
Seq 33 CNVNHKPSNT KVDKKV--EP KSS
41 CH1-5.6 numbered ---ASTKGPS VCPLAPSSKS --TSGGTACL GCSVKDYFPE PVTV-SW--- -using the range of N---SGALT SG-VHTCPAV LQS-SGLYSL WSVVTVPSSS LGT--Q-TYI
Seq 33 CNVNHKPSNT KVDKKV--EP KSS
42 IGG1-HINGE numbered -D--KTHTCP PCP
from 231 to 243 43 IgGl-HINGE-EE -E--KTHTCP ECP
numbered using the range of Seq 42 numbered using the range of Seq 42 45 CH2-WINTER numbered ---APEAAGA PSVFLFPPKP KDTLMI-SRT PEVTCVVVDV
SHEDPEVKFN
using the range of WYV--DG--V EVH-NAKTKP REEQYN---- STYRVVSVLT
VLHQDWLNGK
Seq 13 EYKCKVSNKA LPAPIEKTI- SKAK--46 C53-CW numbered G-QPREPQVC TLPPSREE-- MTKNQVSLWC LVKGFYPS-D IAV--EWES-using the range of NG--QPENNY KTTPPVLDS- D--GSFFLYS KLTVDKSRWQ
QGNVFSCSVM
Seq 18 ("knob") HEALHNHYTQ KSLSLSPGK
47 CH3-CSAV numbered G-QPREPQVY TLPPCREE-- MTKNQVSLSC AVKGFYPS-D IAV--EWES-using the range of NG--QPENNY KTTPPvLDS- D--GSFELVS KLTVDKSRWQ
QGNVFSCSVM
Seq 18 ("hole") HEALHNHYTQ KSLSLSPGK
48 CH3-E numbered G-QPREPQVY TLPPSREE-- MTKNQVSLTC EVKGFYPS-D IAV--EWES-using the range of NG--QPENNY KTTPPVLDS- D--GSFFLYS KLTVDKSRWQ
QGNVFSCSVM
Seq 18 HEALHNHYTQ KSLSLSPGK
49 CH3-R numbered G-QPREPQVY TLPPSREE-- MTKNQVSLTC LVKGFYPS-D IAV--EWES-using the range of NG--QPENNY KTTPPVLDS- D--GSFFLYS RLTVDKSRWQ
QGNVFSCSVM
Seq 18 HEALHNHYTQ KSLSLSPGK
50 C5-VH numbered from QVQLVQSGAE VKKPGSSVKV SCKASGGTFS SYAISWVRQA
PGQGLEWMGG
1 to 113 with IIPIFGTANY AQKFQGRVTI TADESTSTAY MELSSLRSED
TAVYYCAGGS
insertions 52A, ITGTTPFDYW GRGTLVTVSS
82A, 82B, 82C, 100A, 100B, 100C

PGKGLEWVAR
numbered using the INPNNGDTFY TQKFKGRFTI SRDNAKNSLY LQMNSLRAED
TAVYYCARRD
range of Seq 50 YFGAM--DYW GQGTLVTVSS

- 102 -52 C5-VL numbered from AIQLTQSPSS LTASVGDRVT ITCRASQFAS NDVGWYQQKP
GKAPKLLIYA
1 to 107 ASSLQSGVPP RFSGSGSGTE FTFTISSLQP EDFATYYCLQ
DYTYPLTFGQ
GTKVEIK

GKAPKLLIKS
numbered from 1 to ASLAISGVPS RFSGSGSGTD FTLTISSLQP EDFATYYCQQ
SNSWPNTFGG

54 SegID:1,42,45,18 ASTKGPS VFPLAPSSKS TSGGTAAL GCLVKDYFPE PVTV-SW
numbered from 111 N---SGALT SG-VHTFPAV LQS-SGLYSL SSVVTVPSSS LGT--Q-TYI
to 478 with CNVNHKPSNT KVDKKV--EP KSC-D--KTH TCPPCPAPEA
AGAPSVFLFP
insertions 162A PKPKDTLMI- SRTPEVTCVV VDVSHEDPEV KFNWYV--DG --VEVH-NAK
, NKALPAPIEK
, , , TI-SKAKG-Q PREPQVYTLP PSREE--MTK NQVSLTCLVK GFYPS-DIAV --302A, 316A, 316B, EWES-NG- -QPENNYKTT PPVLDS-D-- GSFFLYSKLT VDKSRWQQGN
398A) VFSCSVMHEA LHNHYTQKSL SLSPGK
55 SegID:33,42,45,46 ---ASTKGPS VFPKAPSSKS --TSGGTAAL GCLVKDYFPE PVTV-SW--- -using the numbering N---SGALT SG-VHTFPAV LQS-SGLYSL SSSVTVPSSS LGT--Q-TYI
range of Seq 54 CNVNHKPSNT KVDKKV--EP KSC-D--KTH TCPPCPAPEA
AGAPSVFLFP
PKPKDTLMI- SRTPEVTCVV VDVSHEDPEV KFNWYV--DG --VEVH-NAK
TKPREEQYN- ---STYRVVS VLTVLHQDWL NGKEYKCKVS NKALPAPIEK
TI-SKAKG-Q PREPQVCTLP PSREE--MTK NQVSLWCLVK GFYPS-DIAV --EWES-NG- -QPENNYKTT PPVLDS-D-- GSFFLYSKLT VDKSRWQQGN
VFSCSVMHEA LHNHYTQKSL SLSPGK
56 SegIb:1,42,45,46 ---ASTKGPS VFPLAPSSKS --TSGGTAAL GCLVKDYFPE PVTV-SW
using the numbering N---SGALT SG-VHTFPAV LQS-SGLYSL SSVVTVPSSS LGT--Q-TYI
range of Seq 54 CNVNHKPSNT KVDKKV--EP KSC-D--KTH TCPPCPAPEA
AGAPSVFLFP
PKPKDTLMI- SRTPEVTCVV VDVSHEDPEV KFNWYV--DG --VEVH-NAK
TKPREEQYN- ---STYRVVS VLTVLHQDWL NGKEYKCKVS NKALPAPIEK
TI-SKAKG-Q PREPQVCTLP PSREE--MTK NQVSLWCLVK GFYPS-DIAV --EWES-NG- -QPENNYKTT PPVLDS-D-- GSFFLYSKLT VDKSRWQQGN
VFSCSVMHEA LHNHYTQKSL SLSPGK
57 Seq1D:33,44,45,49 ASTKGPS VFPKAPSSKS --TSGGTAAL GCLVKDYFPE PVTV-SW--- -using the numbering N SGALT SG-VHTFPAV LQS-SGLYSL SSSVTVPSSS LGT--Q-TYI
range of Seq 54 CNVNHKPSNT KVDKKV--EP KSC-R--KTH TCPRCPAPEA
AGAPSVFLFP
PKPKDTLMI- SRTPEVTCVV VDVSHEDPEV KFNWYV--DG --VEVH-NAK
TKPREEQYN- ---STYRVVS VLTVLHQDWL NGKEYKCKVS NKALPAPIEK
TI-SKAKG-Q PREPQVYTLP PSREE--MTK NQVSLTCLVK GFYPS-DIAV --EWES-NG- -QPENNYKTT PPVLDS-D-- GSFFLYSRLT VDKSRWQQGN
VFSCSVMHEA LHNHYTQKSL SLSPGK
58 SeqID:1,44,45,49 ASTKGPS VFPLAPSSKS -TSGGTAAL GCLVKDYFPE PVTV SW
using the numbering N- -SGALT SG-VHTFPAV LQS-SGLYSL SSVVTVPSSS LGT Q TYI
range of Seq 54 CNVNHKPSNT KVDKKV--EP KSC-R--KTH TCPRCPAPEA
AGAPSVFLFP
PKPKDTLMI- SRTPEVTCVV VDVSHEDPEV KENWYV--DG --VEVH-NAK
TKPREEQYN- ---STYRVVS VLTVLHQDWL NGKEYKCKVS NKALPAPIEK
TI-SKAKG-Q PREPQVYTLP PSREE--MTK NQVSLTCLVK GFYPS-DIAV --EWES-NG- -QPENNYKTT PPVLDS-D-- GSFFLYSRLT VDKSRWQQGN
VFSCSVMHEA LHNHYTQKSL SLSPGK
59 Seq1D:34,42,45,47 ---ASTKGPS VFPEAPSSKS --TSGGTAAL GCLVKDYFPE PVTV-SW
using the numbering N---SGALT SG-VHTFPAV LQS-SGLYSL GSVVTVPSSS LGT Q TYI
range of Seq 54 CNVNHKPSNT KVDKKV--EP KSC-D--KTH TCPPCPAPEA
AGAPSVFLFP
PKPKDTLMI- SRTPEVTCVV VDVSHEDPEV KFNWYV--DG --VEVH-NAK
TKPREEQYN- ---STYRVVS VLTVLHQDWL NGKEYKCKVS NKALPAPIEK
TI-SKAKG-Q PREPQVYTLP PCREE--MTK NQVSLSCAVK GFYPS-DIAV --EWES-NG- -QPENNYKTT PPVLDS-D-- GSFFLVSKLT VDKSRWQQGN
VFSCSVMHEA LHNHYTQKSL SLSPGK
60 Seq1D:1,42,45,47 ---ASTKGPS VFPLAPSSKS --TSGGTAAL GCLVKDYFPE PVTV-SW--- -using the numbering N SGALT SG-VHTFPAV LQS-SGLYSL SSVVTVPSSS LGT--Q-TYI
range of Seq 54 CNVNHKPSNT KVDKKV--EP KSC-D--KTH TCPPCPAPEA
AGAPSVFLFP
PKPKDTLMI- SRTPEVTCVV VDVSHEDPEV KENWYV--DG --VEVH-NAK
TKPREEQYN- ---STYRVVS VLTVLHQDWL NGKEYKCKVS NKALPAPIEK
TI-SKAKG-Q PREPQVYTLP PCREE--MTK NQVSLSCAVK GFYPS-DIAV

A

- 103 -EWES-NG- -QPENNYKTT PPVLDS-D-- GSFFLVSKLT VDKSRWQQGN
VFSCSVMHEA LHNHYTQKSL SLSPGK
61 SeqID:35,42,45,47 ---ASTKGPS VFPLAPSSKS --TSGGTAAL GCLVKDYFPE
PVTV-SW--- -using the numbering N---SGALT SG-VHTCPAV LQS-SGLYSL SSIVTVPSSS LGT--Q-TYI
range of Seq 54 CNVNHKPSNT KVDDKV--EP KSS-D--KTH TCPPCPAPEA
AGAPSVFLFP
PKPKDTLMI- SRTPEVTCVV VDVSHEDPEV KFNWYV--DG --VEVH-NAK
TKPREEQYN- ---STYRVVS VLIVLEQDWL NGKEYKCKVS NKALPAPIEK
TI-SKAKG-Q PREPQVYTLP PCREE--MTK NQVSLSCAVK GFYPS-DIAV --EWES-NG- -QPENNYKTT PPVLDS-D-- GSFFLVSKLT VDKSRWQQGN
VFSCSVMHEA LHNHYTQKSL SLSPGK
62 SeqID:36,42,45,47 ---ASTKGPS VCPLAPSSKS --TSGGTAAL GCLVEDYFPE
PVTV-SW--- -using the numbering N---SGALT SG-VHTFPAV LQS-SGLYSL SSVVTVPSSS LGT--Q-TYI
range of Seq 54 CNVNHKPSNT KVDKKV--EP KSS-D--KTH TCPPCPAPEA
AGAPSVFLFP
PKPKDTLMI- SRTPEVTCVV VDVSHEDPEV KFNWYV--DG --VEVH-NAK
TKPREEQYN- ---STYRVVS VLTVLHQDWL NGKEYKCKVS NKALPAPIEK
TI-SKAKG-Q PREPQVYTLP PCREE--MTK NQVSLSCAVK GFYPS-DIAV --EWES-NG- -QPENNYKTT PPVLDS-D-- GSFFLVSKLT VDKSRWQQGN
VFSCSVMHEA LHNHYTQKSL SLSPGK
63 SeqID:37,42,45,47 ---ASTKGPS VFPLAPSSKS --TSGGTAAL GCHVKDYFPE
PVTV-SW--- -using the numbering N---SGALT SG-VHTCPAV LDS-SGLYEL SSIVTVPSSS LGT--Q-TYI
range of Seq 54 CNVNHKPSNT KVDKKV--EP KSS-D--KTH TCPPCPAPEA
AGAPSVFLFP
PKPKDTLMI- SRTPEVTCVV VDVSHEDPEV KFNWYV--DG --VEVH-NAK
TKPREEQYN- ---STYRVVS VLTVLHQDWL NGKEYKCKVS NKALPAPIEK
TI-SKAKG-Q PREPQVYTLP PCREE--MTK NQVSLSCAVK GFYPS-DIAV --EWES-NG- -QPENNYKTT PPVLDS-D-- GSFFLVSKLT VDKSRWQQGN
VFSCSVMHEA LHNHYTQKSL SLSPGK
64 SeqID:38,42,45,47 ---ASTKGPS VFPLAPSSKS --TSGGTACL GCLVSDYFPE
PVTV-SW--- -using the numbering N---SGALT SG-VHTFPAV LQS-SGLYEL SSVVTVPSSS LGT--Q-TYI
range of Seq 54 CNVNHKPSNT KVDKKV--EP KSS-D--KTH TCPPCPAPEA
AGAPSVFLFP
PKPKDTLMI- SRTPEVTCVV VDVSHEDPEV KFNWYV--DG --VEVH-NAK
TKPREEQYN- ---STYRVVS VLTVLHQDWL NGKEYKCKVS NKALPAPIEK
TI-SKAKG-Q PREPQVYTLP PCREE--MTK NQVSLSCAVK GFYPS-DIAV --EWES-NG- -QPENNYKTT PPVLDS-D-- GSFFLVSKLT VDKSRWQQGN
VFSCSVMHEA LHNHYTQKSL SLSPGK
65 Sec:JD:39,42,45,47 ---ASTKGPS VFPLAPSSKS --TSGGTAAL GCSVKDYFPE
PVTV-SW--- -using the numbering N---SGALT SG-VHTCPAV LQS-SGLYSL WSVVTVPSSS LGT--Q-TYI
range of Seq 54 CNVNHKPSNT KVDKKV--EP KSS-D--KTH TCPPCPAPEA
AGAPSVFLFP
PKPKDTLMI- SRTPEVTCVV VDVSHEDPEV KENWYV--DG --VEVH-NAK
TKPREEQYN- ---STYRVVS VLTVLHQDWL NGKEYKCKVS NKALPAPIEK
TI-SKAKG-Q PREPQVYTLP PCREE--MTK NQVSLSCAVK GFYPS-DIAV --EWES-NG- -QPENNYKTT PPVLDS-D-- GSFFLVSKLT VDKSRWQQGN
VFSCSVMHEA LHNHYTQKSL SLSPGK
66 SegID:40,42,45,47 ---ASTKGPS VCFLAPSSKS --TSGGTAAL GCSVKDYFPE

using the numbering N---SGALT SG-VHTFPAV LQS-SGLYSL WSVVTVPSSS LGT--Q-TYI
range of Seq 54 CNVNHKPSNT KVDKKV--EP KSS-D--KTH TCPPCPAPEA
AGAPSVFLFP
PKPKDTLMI- SRTPEVTCVV VDVSHEDPEV KFNWYV--DG --VEVH-NAK
TKPREEQYN- ---STYRVVS VLTVLHQDWL NGKEYKCKVS NKALPAPIEK
TI-SKAKG-Q PREPQVYTLP PCREE--MTK NQVSLSCAVK GFYPS-DIAV --EWES-NG- -QPENNYKTT PPVLDS-D-- GSFFLVSKLT VDKSRWQQGN
VFSCSVMHEA LHNHYTQKSL SLSPGK
67 SeqID:41,42,45,47 ---ASTKGPS VCPLAPSSKS --TSGGTACL GCSVKDYFPE
PVTV-SW--- -using the numbering N---SGALT SG-VHTCPAV LQS-SGLYSL WSVVTVPSSS LGT--Q-TYI
range of Seq 54 CNVNHKPSNT KVDKKV--EP KSS-D--KTH TCPPCPAPEA
AGAPSVFLFP
PKPKDTLMI- SRTPEVTCVV VDVSHEDPEV KFNWYV--DG --VEVH-NAK
TKPREEQYN- ---STYRVVS VLTVLHQDWL NGKEYKCKVS NKALPAPIEK
TI-SKAKG-Q PREPQVYTLP PCREE--MTK NQVSLSCAVK GFYPS-DIAV --EWES-NG- -QPENNYKTT PPVLDS-D-- GSFFLVSKLT VDKSRWQQGN
VFSCSVMHEA LHNHYTQKSL SLSPGK
68 SeqID:34,43,45,48 ---ASTKGPS VFPEAPSSKS --TSGGTAAL GCLVKDYFPE
PVTV-SW--- -using the numbering N---SGALT SG-VHTFPAV LQS-SGLYSL GSVVTVPSSS LGT--Q-TYI

- 104 -range of Seq 54 CNVNHKPSNT KVDKKV--EP KSC-E--KTH TCPECPAPEA
AGAPSVFLFP
PKPKDTLMI- SRTPEVTCVV VDVSHEDPEV KFNWYV--DG --VEVH-NAK
TKPREEQYN- ---STYRVVS VLTVLHQDWL NGKEYKCKVS NKALPAPIEK
TI-SKAKG-Q PREPQVYTLP PSREE--MTK NQVSLTCEVK GFYPS-DIAV --EWES-NG- -QPENNYKTT PPVLDS-D-- GSFFLYSKLT VDKSRWQQGN
VFSCSVMHEA LHNHYTQKSL SLSPGK
69 SeqID:1,43,45,48 ---ASTKGPS VFPLAPSSKS --TSGGTAAL GCLVKDYFPE PVTV-SW--- -using the numbering N---SGALT SG-VHTFPAV LQS-SGLYSL SSVVTVPSSS LGT--Q-TYI
range of Seq 54 CNVNHKPSNT KVDKKV--EP KSC-E--KTH TCPECPAPEA
AGAPSVFLFP
PKPKDTLMI- SRTPEVTCVV VDVSHEDPEV KFNWYV--DG --VEVH-NAK
TKPREEQYN- ---STYRVVS VLTVLHQDWL NGKEYKCKVS NKALPAPIEK
TI-SKAKG-Q PREPQVYTLP PSREE--MTK NQVSLTCEVK GFYPS-DIAV --EWES-NG- -QPENNYKTT PPVLDS-D-- GSFFLYSKLT VDKSRWQQGN
VFSCSVMHEA LHNHYTQKSL SLSPGK
70 SeqID:50,54 QVQLVQSGAE VKKPGSSVKV SCKASGGTFS SYAISWVRQA
PGQGLEWMGG
numbered from 1 to IIPIFGTANY AQKFQGRVTI TADESTSTAY MELSSLRSED
TAVYYCAGGS
478 with insertions ITGTTPFDYW GRGTLVTVSS ASTKGPSVFP LAPSSKS--T SGGTAALGCL
VKDYFPEPVT V-SW----N- --SGALTSG- VHTFPAVLQS -SGLYSLSSV
52A, 82A, 82B, 82C, VTVPSSSLGT --Q-TYICNV NHKPSNTKVD KKV--EPKSC -D--KTHTCP
100A, 100B, 100C, PCPAPEAAGA PSVFLFPPKP KDTLMI-SRT PEVTCVVVDV SHEDPEVKFN
162A, 1623, 162C, WYV--DG--V EVH-NAKTKP REEQYN---- STYRVVSVLT
VLHQDWLNGK
266A, 302A, 316A, EYKCKVSNKA LPAPIEKTI- SKAKG-QPRE PQVYTLPPSR EE--MTKNQV
316B, 398A SLTCLVKGFY PS-DIAV--E WES-NG--QP ENNYKTTPPV LDS-D--GSF
FLYSKLTVDK SRWQQGNVFS CSVMHEALHN HYTQKSLSLS PGK
71 SeqID:50,59 using QVQLVQSGAE VKKPGSSVKV SCKASGGTFS SYAISWVRQA
PGQGLEWMGG
the numbering range IIPIFGTANY AQKFQGRVTI TADESTSTAY MELSSLRSED TAVYYCAGGS
of Seq 70 ITGTTPFDYW GRGTLVTVSS ASTKGPSVFP EAPSSKS--T
SGGTAALGCL
VKDYFPEPVT V SW -N
SGALTSG- VHTFPAVLQS -SGLYSLGSV
VTVPSSSLGT --Q-TYICNV NHKPSNTKVD KKV--EPKSC -D--KTHTCP
PCPAPEAAGA PSVFLFPPKP KDTLMI-SRT PEVTCVVVDV SHEDPEVKFN
WYV--DG--V EVH-NAKTKP REEQYN
STYRVVSVLT VLHQDWLNGK
EYKCKVSNKA LPAPIEKTI- SKAKG-QPRE PQVYTLPPCR EE--MTKNQV
SLSCAVKGFY PS-DIAV--E WES-NG--QP ENNYKTTPPV LDS-D--GSF
FLVSKLTVDK SRWQQGNVFS CSVMHEALHN HYTQKSLSLS PGK
72 SegID:50,60 using QVQLVQSGAE VKKPGSSVKV SCKASGGTFS SYAISWVRQA
PGQGLEWMGG
the numbering range IIPIFGTANY AQKFQGRVTI TADESTSTAY MELSSLRSED TAVYYCAGGS
of Seq 70 ITGTTPFDYW GRGTLVTVSS ASTKGPSVFP LAPSSKS--T
SGGTAALGCL
VKDYFPEPVT V SW -N
SGALTSG- VHTFPAVLQS -SGLYSLSSV
VTVPSSSLGT --Q-TYICNV NHKPSNTKVD KKV--EPKSC -D--KTHTCP
PCPAPEAAGA PSVFLFPPKP KDTLMI-SRT PEVTCVVVDV SHEDPEVKFN
WYV--DG--V EVH-NAKTKP REEQYN
STYRVVSVLT VLHQDWLNGK
EYKCKVSNKA LPAPIEKTI- SKAKG-QPRE PQVYTLPPCR EE--MTKNQV
SLSCAVKGFY PS-DIAV--E WES-NG--QP ENNYKTTPPV LDS-D--GSF
FLVSKLTVDK SRWQQGNVFS CSVMHEALHN HYTQKSLSLS PGK
73 SeqID:50,68 using QVQLVQSGAE VKKPGSSVKV SCKASGGTFS SYAISWVRQA
PGQGLEWMGG
the numbering range IIPIFGTANY AQKFQGRVTI TADESTSTAY MELSSLRSED TAVYYCAGGS
of Seq 70 ITGTTPFDYW GRGTLVTVSS ASTKGPSVFP EAPSSKS--T
SGGTAALGCL
VKDYFPEPVT V-SW----N- --SGALTSG- VHTFPAVLQS -SGLYSLGSV
VTVPSSSLGT --Q-TYICNV NHKPSNTKVD KKV--EPKSC -E--KTHTCP
ECPAPEAAGA PSVFLFPPKP KDTLMI-SRT PEVTCVVVDV SHEDPEVKFN
WYV--DG--V EVH-NAKTKP REEQYN---- STYRVVSVLT VLHQDWLNGK
EYKCKVSNKA LPAPIEKTI- SKAKG-QPRE PQVYTLPPSR EE--MTKNQV
SLTCEVKGFY PS-DIAV--E WES-NG--QP ENNYKTTPPV LDS-D--GSF
FLYSKLTVDK SRWQQGNVFS CSVMHEALHN HYTQKSLSLS PGK
74 SeqID:50,69 using QVQLVQSGAE VKKPGSSVKV SCKASGGTFS SYAISWVRQA
PGQGLEWMGG
the numbering range IIPIFGTANY AQKFQGRVTI TADESTSTAY MELSSLRSED TAVYYCAGGS
of Se 70 ITGTTPFDYW GRGTLVTVSS ASTKGPSVFP LAPSSKS--T
SGGTAALGCL
q VKDYFPEPVT V SW -N
SGALTSG- VHTFPAVLQS -SGLYSLSSV
VTVPSSSLGT --Q-TYICNV NHKPSNTKVD KKV--EPKSC -E--KTHTCP
ECPAPEAAGA PSVFLFPPKP KDTLMI-SRT PEVTCVVVDV SHEDPEVKFN

- 105 -WYV--DG--V EVH-NAKTKP REEQYN
STYRVVSVLT VLHQDWLNGK
EYKCKVSNKA LPAPIEKTI- SKAKG-QPRE PQVYTLPPSR EE--MTKNQV
SLTCEVKGFY PS-DIAV--E WES-NG--QP ENNYKTTPPV LDS D GSF
FLYSKLTVDK SRWQQGNVFS CSVMHEALHN HYTQKSLSLS PGK
75 SeqID:51,54 using EVQLVESGGG LVQPGGSLRL SCAASGYSFT AYFMNWVRQA
PGKGLEWVAR
the numbering range INPNNGDTFY TQKFKGRFTI SRDNAKNSLY LQMNSLRAED TAVYYCARRD
of Seq 70 YFGAM--DYW GQGTLVTVSS ASTKGPSVFP LAPSSKS--T
SGGTAALGCL
VKDYFPEPVT V-SW N-SGALTSG- VHTFPAVLQS -SGLYSLSSV
VTVPSSSLGT --Q-TYICNV NHKPSNTKVD KKV--EPKSC -D--KTHTCP
PCPAPEAAGA PSVFLFPPKP KDTLMI-SRT PEVTCVVVDV SHEDPEVKFN
WYV--DG--V EVH-NAKTKP REEQYN
STYRVVSVLT VLHQDWLNGK
EYKCKVSNKA LPAPIEKTI- SKAKG-QPRE PQVYTLPPSR EE--MTKNQV
SLTCLVKGFY PS-DIAV--E WES-NG--QP ENNYKTTPPV LDS-D--GSF
FLYSKLTVDK SRWQQGNVFS CSVMHEALHN HYTQKSLSLS PGK
76 SeqID:52,9 numbered AIQLTQSPSS LTASVGDRVT ITCRASQFAS NDVGWYQQKP
GKAPKLLIYA
from position 1 to ASSLQSGVPP RFSGSGSGTE FTFTISSLQP EDFATYYCLQ
DYTYPLTFGQ

PREAKVQWKV
DNALQSGNSQ ESVTEQDSKD STYSLSSTLT LSKADYEKHK VYACEVTHQG
LSSPVTKSFN RGEC-77 SeqID:52,25 using AIQLTQSPSS LTASVGDRVT ITCRASQFAS NDVGWYQQKP
GKAPKLLIYA
the numbering range ASSLQSGVPP RFSGSGSGTE FTFTISSLQP EDFATYYCLQ DYTYPLTFGQ
of Seq 76 GTKVEIKRTV AAPSVFIFPP SDEQLKSGTA SVSCLLNNFY
PREAKVQWKV
DNALQSGNSQ ESVTEQDSKD STYSLKSTLT LSKADYEKHK VYACEVTHQG
LSSPVTKSFN RGEC-78 SegID:53,9 using DIQMTQSPSS LSASVGDRVT ITCRASQTIS NNLHWYQQKP
GKAPKLLIKS
the numbering range ASLAISGVPS RFSGSGSGTD FTLTISSLQP EDFATYYCQQ SNSWPNTFGG
of Seq 76 GTKVEIKRTV AAPSVFIFPP SDEQLKSGTA SVVCLLNNFY
PREAKVQWKV
DNALQSGNSQ ESVTEQDSKD STYSLSSTLT LSKADYEKHK VYACEVTHQG
LSSPVTKSFN RGEC-79 IgG2 Hinge using -CC-V-E-CP PCP
the numbering range of Seq 42 80 IgG3 Hinge from 231 -LGDTTHTCP RCPEPKSCDT PPPCPRCPEP KSCDTPPPCP
RCPEPKSCDT
to 243 with PPPCPRCP
insertions 241A
through 241Z, then insertions 241AA
through 241SS
81 Ig04 Hinge using PPCP SCP
the numbering range of Seq 42 82 IgC1 CH3 alternate G-QPREPQVY TLPPSREE-- MTKNQVSLTC EVKGFYPS-D IAV--EWES-allotype numbered NG--QPENNY KTTPPVLDS- D--GSFFLYS KLTVDKSRWQ
QGNVFSCSVM
from 361 to 478 HEALHNHYTQ KSLSLSPGK
with E391 83 IgG1 CH3 alternate G-QPREPQVY TLPPSREE-- MTKNQVSLTC LVKGFYPS-D IAV--EWES-allotype numbered NG--QPENNY KTTPPVLDS- D--GSFFLYS RLTVDKSRWQ
QGNVFSCSVM
from 361 to 478 HEALHNHYTQ KSLSLSPGK
with R441 84 IgG2 CH3 alternate G-QPREPQVY TLPPSREE-- MTKNQVSLTC LVKGFYPS-D ISV--EWES-allotype numbered NG--QPENNY KTTPPMLDS- D--GSFFLYS RLTVDKSRWQ
QGNVFSCSVM
from 361 to 478 HEALHNHYTQ KSLSLSPGK
with R441 85 IgG2 CH3 alternate G-QPREPQVY TLPPSREE-- MTKNQVSLTC EVKGFYPS-D ISV--EWES-allotype numbered NG--QPENNY KTTPPMLDS- D--GSFFLYS KLTVDKSRWQ
QGNVFSCSVM
from 361 to 478 HEALHNHYTQ KSLSLSPGK
with E391

- 106 -86 CL-S176D using the RTV
AAPSVFIFPP SDEQLKSGTA SVVCLLNNFY PREAKVQWKV
numbering range of DNALQSGNSQ ESVTEQDSKD STYSLDSTLT LSKADYEKHK
VYACEVTHQG
Seq 9 LSSPVTKSFN RGEC
87 CL-S176K using the RTV
AAPSVFIFPP SDEQLKSGTA SVVCLLNNFY PREAKVQWKV
numbering range of DNALQSGNSQ ESVTEQDSKD STYSLESTLT LSKADYEKHK
VYACEVTHQG
Seq 9 LSSPVTKSFN RGEC
88 CL-Deconvolute-05 AIQLTQSPSS LTASVGDRVT ITCRASQFAS NDVGWYQQKP
GKAPKLLIYA
arm using the ASSLQSGVPP RFSGSGSGTE FTFTISSLQP EDFATYYCLQ
DYTYPLTFGQ
numbering range of GTKVEIKRTV AAPSVFIFPP SDEQLKSGTA SVVCLLNNFY
PREAKVQWKV
Seq 76 DNALQSGNSQ ESVTEQDSKD STYSLESTLT LSKADYEKHK
VYACEVTHQG
LSSPVTKSFN RGEC-89 CH1-L124K numbered ---ASTKGPS VFPEAPSSKS --TSGGTAAL GCLVKDYFPE PVTV-SW--- -using the range of N---SGALT SG-VHTFPAV LQS-SGLYSL SSVVTVPSSS LGT--Q-TYI
Seq 33 CNVNHKPSNT KVDKKV--EP KSC
90 CH1-L1245 numbered ---ASTKGPS VFPEAPSSKS --TSGGTAAL GCLVKDYFPE PVTV-SW--- -using the range of N---SGALT SG-VHTFPAV LQS-SGLYSL SSVVTVPSSS LGT--Q-TYI
Seq 33 CNVNHKPSNT KVDKKV--EP KSC
91 CH-L124K-Knob (aka ---ASTKGPS VFPEAPSSKS --TSGGTAAL GCLVKDYFPE PVTV-SW
Seq 89,42,45,46) N---SGALT SG-VHTFPAV LQS-SGLYSL SSVVTVPSSS LGT--Q-TYI
numbered using the CNVNHKPSNT KVDKKV--EP KSC-D--KTH TCPPCPAPEA
AGAPSVFLFP
range of Seq 54 PKPKDTLMI- SRTPEVTCVV VDVSHEDPEV KENWYV--DG --VEVH-NAK
TKPREEQYN- ---STYRVVS VLTVLHQDWL NGKEYKCKVS NKALPAPIEK
TI-SKAKG-Q PREPQVCTLP PSREE--MTK NQVSLWCLVK GFYPS-DIAV --EWES-NG- -QPENNYKTT PPVLDS-D-- GSFFLYSKLT VDKSRWQQGN
VFSCSVMHEA LHNHYTQKSL SLSPGK
92 CH-L124E-Hole (aka ---ASTKGPS VFPEAPSSKS --TSGGTAAL GCLVKDYFPE PVTV-SW--- -Seq 90,42,45,47) N---SGALT SG-VHTFPAV LQS-SGLYSL SSVVTVPSSS LGT--Q-TYI
numbered using the CNVNHKPSNT KVDKKV--EP KSC-D--KTH TCPPCPAPEA
AGAPSVFLFP
range of Seq 54 PKPKDTLMI- SRTPEVTCVV VDVSHEDPEV KFNWYV--DG --VEVH-NAK
TKPREEQYN- ---STYRVVS VLTVLHQDWL NGKEYKCKVS NKALPAPIEK
TI-SKAKG-Q PREPQVYTLP PCREE--MTK NQVSLSCAVK GFYPS-DIAV --EWES-NG- -QPENNYKTT PPVLDS-D-- GSFFLVSKLT VDKSRWQQGN
VFSCSVMHEA LHNHYTQKSL SLSPGK
93 HC-05_Deconvolute QVQLVQSGAE VKKPGSSVKV SCKASGGTFS SYAISWVRQA
PGQGLEWMGG
Si rev (aka Seq IIPIFGTANY AQKFQGRVTI TADESTSTAY MELSSLRSED
TAVYYCAGGS
50,92 or ITGTTPFDYW GRGTLVTVSS ASTKGPSVFP EAPSSKS--T
SGGTAALGCL
50 90 42 45 47) VKDYFPEPVT V-SW----N- --SGALTSG- VHTFPAVLQS -SGLYSLSSV
,,,, VTVPSSSLGT --Q-TYICNV NHKPSNTKVD KKV--EPKSC -D--KTHTCP
using the numbering PCPAPEAAGA PSVFLFPPKP KDTLMI-SRT PEVTCVVVDV SHEDPEVKFN
range of Seq 70 WYV--DG--V EVH-NAKTKP REEQYN
STYRVVSVLT VLHQDWLNGK
EYKCKVSNKA LPAPIEKTI- SKAKG-QPRE PQVYTLPPCR EE--MTKNQV
SLSCAVKGFY PS-DIAV--E WES-NG--QP ENNYKTTPPV LDS-D--GSF
FLVSKLTVDK SRWQQGNVFS CSVMHEALHN HYTQKSLSLS PCK
94 LC-05-T1 (aka Seq AIQLTQSPSS LTASVGDRVT ITCRASQFAS NDVGWYQQKP
GKAPKLLIYA
52,26) using the ASSLQSGVPP RFSGSGSGTE FTFTISSLQP EDFATYYCLQ
DYTYPLTFGQ
numbering range of GTKVEIKRTV AAPSVFIFPP SDKQLKSGTA SVVCILNNFY
PREAKVQWKV
Se 76 DNALQSGNSQ ESVTEQDSKD STYSLCSTLT LSKADYEKHK
VYACEVTHQG
q LSSPVTKSFN RGES-95 LC-05-T2 (aka Seq AIQLTQSPSS LTASVGDRVT ITCRASQFAS NDVGWYQQKP
GKAPKLLIYA
52,27) using the ASSLQSGVPP RFSGSGSGTE FTFTISSLQP EDFATYYCLQ
DYTYPLTFGQ
numbering range of GTKVEIKRTV AAPSVFIFPP SDCQLKSGTA HVVCLLNNFY
PREAKVQWKV
Se 76 DNALQSGNSQ ESVTEQDSKD STYSLSSTLT LSKADYEKHK
VYACEVTHQG
q LSSPVTKSFN ROES-96 LC-05-T3 (aka Seq AIQLTQSPSS LTASVGDRVT ITCRASQFAS NDVGWYQQKP
GKAPKLLIYA
52,28) using the ASSLQSGVPP RFSGSGSGTE FTFTISSLQP EDFATYYCLQ
DYTYPLTFGQ
numbering range of GTKVEIKRTV AAPSVFIFPP SDEQLKSGTA HVVCILNNFY
PREAKVQWKV
Se 76 DNALQSGNSQ ESVTEQDSKD STYSLCSTLT LSKADYEKHK
VYACEVTHQG
q LSSPVTKSFN RGES-97 LC-05-T4 (aka Seq AIQLTQSPSS LTASVGDRVT ITCRASQFAS NDVGWYQQKP
GKAPKLLIYA

- 107 -52,29) using the ASSLQSGVPP RFSGSGSGTE FTFTISSLQP EDFATYYCLQ
DYTYPLTFGQ
numbering range of GTKVEIKRTV AAPSVCIFPP SDEQLKSGTA HVVCLLNNFY
PREAKVQWKV
Seq 76 DNALQSGNSQ ESVTEQDSKD STYSLSSTLT LSKADYEKHK
VYACEVTHQG
LSSPVTKSFN RGES-98 LC-05-T9 (aka Seq AIQLTQSPSS LTASVGDRVT ITCRASQFAS NDVGWYQQKP
GKAPKLLIYA
52,30) using the ASSLQSGVPP RFSGSGSGTE FTFTISSLQP EDFATYYCLQ
DYTYPLTFGQ
numbering range of GTKVEIKRTV AAPSVFIFPP SDEQLKSGTA DVSCLLNNFY
PREAKVQWKV
Se 76 DNALQSGNSQ ESVTEQDSKD STYSLCSSLT LSKADYEKHK
VYACEVTHQG
q LSSPVTKSFN RGES-99 HC-05-T1 (aka Seq QVQLVQSGAE VKKPGSSVKV SCKASGGTFS SYAISWVRQA
PGQGLEWMGG
50,61) using the IIPIFGTANY AQKFQGRVTI TADESTSTAY MELSSLRSED
TAVYYCAGGS
numbering range of ITGTTPFDYW GRGTLVTVSS ASTKGPSVFP LAPSSKS--T
SGGTAALGCL
Seq 70 VKDYFPEPVT V SW SGALTSG- VHTCPAVLQS -SGLYSLSSI
VTVPSSSLGT --Q-TYICNV NHKPSNTKVD DKV--EPKSS -D--KTETCP
PCPAPEAAGA PSVFLFPPKP KDTLMI-SRT PEVTCVVVDV SHEDPEVKFN
WYV--DG--V EVH-NAKTKP REEQYN---- STYRVVSVLT VLHQDWLNGK
EYKCKVSNKA LPAPIEKTI- SKAKG-QPRE PQVYTLPPCR EE--MTKNQV
SLSCAVKGFY PS-DIAV--E WES-NG--QP ENNYKTTPPV LDS-D--GSF
FLVSKLTVDK SRWQQGNVFS CSVMHEALHN HYTQKSLSLS PGK
100 HC-05-T2 (aka Seq QVQLVQSGAE VKKPGSSVKV SCKASGGTFS SYAISWVRQA
PGQGLEWMGG
50,62) using the IIPIFGTANY AQKFQGRVTI TADESTSTAY MELSSLRSED
TAVYYCAGGS
numbering range of ITGTTPFDYW GRGTLVTVSS ASTKGPSVCP LAPSSKS--T
SGGTAALGCL
Se 70 VEDYFPEPVT V-SW----N- --SGALTSG- VHTFPAVLQS -SGLYSLSSV
q VTVPSSSLGT --Q-TYICNV NHKPSNTKVD KKV--EPKSS -D--KTHTCP
PCPAPEAAGA PSVFLFPPKP KDTLMI-SRT PEVTCVVVDV SHEDPEVKFN
WYV--DG--V EVH-NAKTKP REEQYN---- STYRVVSVLT VLHQDWLNGK
EYKCKVSNKA LPAPIEKTI- SKAKG-QPRE PQVYTLPPCR EE--MTKNQV
SLSCAVKGFY PS-DIAV--E WES-NG--QP ENNYKTTPPV LDS-D--GSF
FLVSKLTVDK SRWQQGNVFS CSVMHEALHN HYTQKSLSLS PGK
101 HC-05-T3 (aka Seq QVQLVQSGAE VKKPGSSVKV SCKASGGTFS SYAISWVRQA
PGQGLEWMGG
50,63) using the IIPIFGTANY AQKFQGRVTI TADESTSTAY MELSSLRSED
TAVYYCAGGS
numbering range of ITGTTPFDYW GRGTLVTVSS ASTKGPSVFP LAPSSKS--T
SGGTAALGCH
Se 70 VKDYFPEPVT V SW -N SGALTSG- VHTCPAVLDS -SGLYELSSI
q VTVPSSSLGT --Q-TYICNV NHKPSNTKVD KKV--EPKSS -D--KTHTCP
PCPAPEAAGA PSVFLFPPKP KDTLMI-SRT PEVTCVVVDV SHEDPEVKFN
WYV--DG--V EVH-NAKTKP REEQYN---- STYRVVSVLT VLHQDWLNGK
EYKCKVSNKA LPAPIEKTI- SKAKG-QPRE PQVYTLPPCR EE--MTKNQV
SLSCAVKGFY PS-DIAV--E WES-NG--QP ENNYKTTPPV LDS-D--GSF
FLVSKLTVDK SRWQQGNVFS CSVMHEALHN HYTQKSLSLS PGK
102 HC-05-T4 (aka Seq QVQLVQSGAE VKKPGSSVKV SCKASGGTFS SYAISWVRQA
PGQGLEWMGG
50,64) using the IIPIFGTANY AQKFQGRVTI TADESTSTAY MELSSLRSED
TAVYYCAGGS
numbering range of ITGTTPFDYW GRGTLVTVSS ASTKGPSVFP LAPSSKS--T
SGGTACLGCL
Se 70 VSDYFPEPVT V-SW----N- --SGALTSG- VHTFRAVLQS -SGLYELSSV
g VTVPSSSLGT --Q-TYICNV NHKPSNTKVD KKV--EPKSS -D--KTHTCP
PCPAPEAAGA PSVFLFPPKP KDTLMI-SRT PEVTCVVVDV SHEDPEVKFN
WYV--DG--V EVH-NAKTKP REEQYN---- STYRVVSVLT VLHQDWLNGK
EYKCKVSNKA LPAPIEKTI- SKAKG-QPRE PQVYTLPPCR EE--MTKNQV
SLSCAVKGFY PS-DIAV--E WES-NG--QP ENNYKTTPPV LDS-D--GSF
FLVSKLTVDK SRWQQGNVFS CSVMHEALHN HYTQKSLSLS PGK
103 HC-05-T9 (aka Seq QVQLVQSGAE VKKPGSSVKV SCKASGGTFS SYAISWVRQA
PGQGLEWMGG
50,65) using the IIPIFGTANY AQKFQGRVTI TADESTSTAY MELSSLRSED
TAVYYCAGGS
numbering range of ITGTTPFDYW GRGTLVTVSS ASTKGPSVFP LAPSSKS--T
SGGTAALGCS
Se 70 VKDYFPEPVT V-SW----N- --SGALTSG- VHTCPAVLQS -SGLYSLWSV
q VTVPSSSLGT --Q-TYICNV NHKPSNTKVD KKV--EPKSS -D--KTETCP
PCPAPEAAGA PSVFLFPPKP KDTLMI-SRT PEVTCVVVDV SHEDPEVKFN
WYV--DG--V EVH-NAKTKP REEQYN---- STYRVVSVLT VLHQDWLNGK
EYKCKVSNKA LPAPIEKTI- SKAKG-QPRE PQVYTLPPCR EE--MTKNQV
SLSCAVKGFY PS-DIAV--E WES-NG--QP ENNYKTTPPV LDS-D--GSF
FLVSKLTVDK SRWQQGNVFS CSVMHEALHN HYTQKSLSLS PGK

- 108 -GTTCTCTGAGGTTCAGCTGCAGCAGTCTGGACCTGAGCTGGTGAAGCC
TGGGGCTTCAATGAAGATATCCTGCAAGACTTCTGGTTACTCATTTACTG
CCTACTTTATGAACTGGGTGAAGCAGAGCCATGGAAAGAGCCTTGAGTG
GATTGGACGTATTAATCCCAACAATGGTGACACTTTCTACACCCAGAAGT
TCAAGGGCAAGGCCACATTGACTGTAGACAAATCCTCTAACACAGCCCA
CATGGAACTCCTGAGCCTGACATCTGAGGACTCTGCAATCTATTATTGT
GGAAGAAGGGATTATTTCGGGGCTATGGACTACTGGGGTCAAGGAACC
TCAGTCACCGTCTCCTCA
105 TA-1 VH (leader, CDRs MGWSC/FLFLLSVTVGVFSEVQLQQSGPELVKPGASMKISCKTSGYSFTAY
underlined and defined by FMNVVVKQSHGKSLEWIGRINPNNGDTFYTQKFKGKATLTVDKSSNTAHME
Kabat) LLSLTSEDSAIYYCGRRDYFGAMDYVVGQGTSVTVSS

CACACCTCAGATACTTGGACTTATGCTTTTTTGGATTTCAGC
CTCCAGAGGTGCTATTGTGCTAATTCAGTCTCCAGCCACCCTGTCTGTG
ACTCCAGGAGATAGCGTCAGTCTTTCCTGCAGGGCCAGCCAAACTATTA
GTAACAACCTACACTGGTATCAACAAAAATCACATGAGTCTCCAAGGCTT
CTCATCAAGTCTGCTTCCCTGGCCATCTCTGGGATCCCCTCCAGGTTCA
GTGGCAGTGGATCAGGGACAGATTTCACTCTCAGTATCAGCAGTGTGGA
GACTGAAGA I -I I I GGAATGTATTTCTGTCAACAGAGTAACAGCTGGCCG
AACACGTTCGGCGGGGGGACCAAGCTGGAAATAAAA
107 TOA-1 VL (leader, CDRs MVFTPQ/LGLMLFW/SASRGAIVLIQSPATLSVTPGDSVSLSCRASQTISNNL
underlined defined by Kabat) HVVYQQKSHESPRLLIKSASLAISGIPSRFSGSGSGTDFTLSISSVETEDFGM
YFCQQSNSWPNTFGGGTKLEIK
108 SEQ ID NO: 5 TOA-1 Al D
GATATTGTGCTAATTCAGTCTCCAGCCACCCTGTCTGTGACTCCAGGAG
VL ATAGCGTCAGTCTTTCCTGCAGGGCCAGCCAAACTATTAGTAACAACCT
ACACTGGTATCAACAAAAATCACATGAGTCTCCAAGGCTTCTCATCAAGT
CTGCTTCCCTGGCCATCTCTGGGATCCCCTCCAGGTTCAGTGGCAGTG
GATCAGGGACAGATTTCACTCTCAGTATCAGCAGTGTGGAGACTGAAGA

GGCGGGGGGACCAAGGTGGAAATAAAA

109 TOA-1 AID VL (leader, DIVLIQSPATLSVTPGDSVSLSCRASQTISNNLHVVYQQKSHESPRLLIKSASL
CDRs underlined defined by AISGIPSRFSGSGSGTDFTLSISSVETEDFGMYFCQQSNSWPNTFGGGTKV
Kabat) EIK

110 huT0A-1 VH v1.0 GAGGTGCAGCTGGTGGAGTCTGGGGGAGGCTTGGTCCAGCCTGGGGG
GTCCCTGAGACTCTCCTGTGCAGCCTCTGGTTACTCATTTACTGCCTACT
TTATGAACTGGGTCCGCCAGGCTCCAGGGAAGGGGCTGGAGTGGGTG
GCCCGTATTAATCCCAACAATGGTGACACTTTCTACACCCAGAAGTTCAA
GGGCCGATTCACCATCTCCAGAGACAACGCCAAGAACTCACTGTATCTG
CAAATGAACAGCCTGAGAGCCGAGGACACCGCTGTGTATTACTGTGCCA
GAAGGGATTATTTCGGGGCTATGGACTACTGGGGTCAAGGAACCTTGGT
CACCGTCTCCTCA

111 huT0A-1 VH v1.1 (CDRs EVQLVESGGGLVQPGGSLRLSCATSGYSFTAYFMNVVVRQAPGKGLEVVVA
underlined defined by AbM) RINPNNGDIFYTQKFKGRFTISVDNAKNSAYLQMNSLRAEDTAVYYCARRD

YFGAMDYVVGQGTLVTVSS

112 huT0A-1 VH v1.1 GAGGTGCAGCTGGTGGAGTCTGGGGGAGGCTTGGTCCAGCCTGGGGG
GTCCCTGAGACTCTCCTGTGCAACCTCTGGTTACTCATTTACTGCCTACT
TTATGAACTGGGTCCGCCAGGCTCCAGGGAAGGGGCTGGAGTGGGTG
GCCCGTATTAATCCCAACAATGGTGACACTTTCTACACCCAGAAGTTCAA
GGGCCGATTCACCATCTCCGTGGACAACGCCAAGAACTCAGCCTATCTG
CAAATGAACAGCCTGAGAGCCGAGGACACCGCTGTGTATTACTGTGCCA
GAAGGGATTATTTCGGGGCTATGGACTACTGGGGTCAAGGAACCTTGGT
CACCGTCTCCTCA

113 huT0A-1 VH v2.0 (CDRs EVQLVQSGAEVKKPGATVKISCKVSGYSFTAYFMNVVVQQAPGKGLEWMG
underlined defined by AbM) RINPNNGDTFYTQKFKGRVTITADTSTDTAYMELSSLRSEDTAVYYCATRDY
FGAMDYWGQGTLVTVSS

114 huT0A-1 VH v2.0 GAGGTCCAGCTGGTGCAGTCTGGGGCTGAGGTGAAGAAGCCTGGGGC
TACAGTGAAAATCTCCTGCAAGGTCTCCGGTTACTCATTTACTGCCTACT
TTATGAACTGGGTGCAACAGGCCCCTGGAAAAGGGCTGGAGTGGATGG
GACGTATTAATCCCAACAATGGTGACACTTTCTACACCCAGAAGTTCAAG

GGCAGAGTCACCATAACCGCTGACACCTCTACAGACACAGCCTACATGG
AGCTGAGCAGCCTGCGCTCTGAGGACACCGCCGTGTATTACTGTGCAA
CAAGGGATTATTTCGGGGCTATGGACTACTGGGGTCAAGGAACCTIGGT
CACCGTCTCCTCA

115 huT0A-1 VL v1.0 (CDRs DIQMTQSPSSLSASVGDRVTITCRASQTISNNLHVVYQQKPGKAPKWYSAS
underlined defined by AbM) LAISGVPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQSNSWPNTFGGGTK

VEIK

116 huT0A-1 VL v1.0 GACATCCAGATGACCCAGTCTCCATCCTCCCTGTCTGCATCTGTCGGAG
ACAGAGTCACCATCACTTGCAGGGCCAGCCAAACTATTAGTAACAACCT
GCACTGGTATCAGCAGAAACCAGGCAAAGCCCCTAAGCTCCTGATCTAT
TCTGCTTCCCTGGCCATCTCTGGAGTCCCATCCCGCTTCAGCGGCAGC
GGATCCGGCACAGATTTCACTCTCACCATCAGCAGCCTGCAACCTGAAG
AIIII GCAACTTACTACTGTCAACAGAGTAACAGCTGGCCCAACACCTTC
GGCGGAGGGACCAAGGTGGAAATAAAA

117 huT0A-1 VL v1.1 (CDRs DIQMTQSPSSLSASVGDRVTITCRASQTISNNLHVVYQQKPGESPKLLIKSAS
underlined defined by AbM) LAISGVPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQSNSWPNTFGGGTK

VEIK

118 huT0A-1 VL v1.1 GACATCCAGATGACCCAGTCTCCATCCTCCCTGTCTGCATCTGTCGGAG
ACAGAGTCACCATCACTTGCAGGGCCAGCCAAACTATTAGTAACAACCT
GCACTGGTATCAGCAGAAACCAGGCGAGTCCCCTAAGCTCCTGATCAA
GTCTGCTTCCCTGGCCATCTCTGGAGTCCCATCCCGCTTCAGCGGCAG
CGGATCCGGCACAGATTTCACTCTCACCATCAGCAGCCTGCAACCTGAA

CGGCGGAGGGACCAAGGTGGAAATAAAA

119 huT0A-1 VL v1.2 (CDRs DIQMTQSPSSLSASVGDRVTITCRASQTISNNLHWYQQKPGEAPKWYSAS
underlined defined by AbM) LAISGVPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQSNSWPNTFGGGTK

VEIK

120 huT0A-1 VL v1.2 GACATCCAGATGACCCAGTCTCCATCCTCCCTGTCTGCATCTGTCGGAG
ACAGAGTCACCATCACTTGCAGGGCCAGCCAAACTATTAGTAACAACCT
GCACTGGTATCAGCAGAAACCAGGCGAGGCCCCTAAGCTCCTGATCTAT
TCTGCTTCCCTGGCCATCTCTGGAGTCCCATCCCGCTTCAGCGGCAGC
GGATCCGGCACAGATTTCACTCTCACCATCAGCAGCCTGCAACCTGAAG
AIIII GCAACTTACTACTGTCAACAGAGTAACAGCTGGCCCAACACCTTC
GGCGGAGGGACCAAGGTGGAAATAAAA

121 huT0A-1 VL v1.3 (CDRs DIQMTQSPSSLSASVGDRVTITCRASQTISNNLHVVYQQKPGKSPKWYSAS
underlined defined by AbM) LAISGVPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQSNSWPNTFGGGTK

VEIK

122 SEQ ID NO: 20 huT0A-1 VL
GACATCCAGATGACCCAGTCTCCATCCTCCCTGTCTGCATCTGTCGGAG
v1.3 ACAGAGTCACCATCACTTGCAGGGCCAGCCAAACTATTAGTAACAACCT
GCACTGGTATCAGCAGAAACCAGGCAAATCCCCTAAGCTCCTGATCTAT
TCTGCTTCCCTGGCCATCTCTGGAGTCCCATCCCGCTTCAGCGGCAGC
GGATCCGGCACAGATTTCACTCTCACCATCAGCAGCCTGCAACCTGAAG
AI I I I GCAACTTACTACTGTCAACAGAGTAACAGCTGGCCCAACACCTTC
GGCGGAGGGACCAAGGTGGAAATAAAA

123 huT0A-1 VL v1.4 GACATCCAGATGACCCAGTCTCCATCCTCCCTGTCTGCATCTGTCGGAG
ACAGAGTCACCATCACTTGCAGGGCCAGCCAAACTATTAGTAACAACCT
GCACTGGTATCAGCAGAAACCAGGCAAAGCCCCTAAGCTCCTGATCAAG
TCTGCTTCCCTGGCCATCTCTGGAGTCCCATCCCGCTTCAGCGGCAGC
GGATCCGGCACAGATTTCACTCTCACCATCAGCAGCCTGCAACCTGAAG
AIIII GCAACTTACTACTGTCAACAGAGTAACAGCTGGCCCAACACCTTC
GGCGGAGGGACCAAGGTGGAAATAAAA

124 huT0A-1 VL v1.5 (CDRs DIQMTQSPSSLSASVGDRVTITCRASQTISNNLHVVYQQKPHKAPKWYSAS
underlined defined by AbM) LAISGVPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQSNSWPNTFGGGTK

VEIK

125 huT0A-1 VL v1.5 GACATCCAGATGACCCAGTCTCCATCCTCCCTGTCTGCATCTGTCGGAG
ACAGAGTCACCATCACTTGCAGGGCCAGCCAAACTATTAGTAACAACCT
GCACTGGTATCAGCAGAAACCACACAAAGCCCCTAAGCTCCTGATCTAT
TCTGCTTCCCTGGCCATCTCTGGAGTCCCATCCCGCTTCAGCGGCAGC

GGATCCGGCACAGATTTCACTCTCACCATCAGCAGCCTGCAACCTGAAG
ATTTTGCAACTTACTACTGTCAACAGAGTAACAGCTGGCCCAACACCTTC
GGCGGAGGGACCAAGGTGGAAATAAAA

126 huT0A-1 VL v1.6 (CDRs DIQMTQSPSSLSASVGDRVTITCRASQTISNNLHVVYQQKPGESPKWYSAS
underlined defined by AbM) LAISGVPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQSNSWPNTFGGGTK

VEIK

127 huT0A-1 VL v1.6 GACATCCAGATGACCCAGTCTCCATCCTCCCTGTCTGCATCTGTCGGAG
ACAGAGTCACCATCACTTGCAGGGCCAGCCAAACTATTAGTAACAACCT
GCACTGGTATCAGCAGAAACCAGGCGAGTCCCCTAAGCTCCTGATCTAT
TCTGCTTCCCTGGCCATCTCTGGAGTCCCATCCCGCTTCAGCGGCAGC
GGATCCGGCACAGATTTCACTCTCACCATCAGCAGCCTGCAACCTGAAG
ATTTTGCAACTTACTACTGTCAACAGAGTAACAGCTGGCCCAACACCTTC
GGCGGAGGGACCAAGGTGGAAATAAAA

128 huT0A-1 VL v1.7 (CDRs DIQMTQSPSSLSASVGDRVTITCRASQTISNNLHVVYQQKPGKSPKLLIKSAS
underlined defined by AbM) LAISGVPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQSNSWPNTFGGGTK

VEIK

129 huT0A-1 VL v1.7 GACATCCAGATGACCCAGTCTCCATCCTCCCTGTCTGCATCTGTCGGAG
ACAGAGTCACCATCACTTGCAGGGCCAGCCAAACTATTAGTAACAACCT
GCACTGGTATCAGCAGAAACCAGGCAAATCCCCTAAGCTCCTGATCAAG
TCTGCTTCCCTGGCCATCTCTGGAGTCCCATCCCGCTTCAGCGGCAGC
GGATCCGGCACAGATTTCACTCTCACCATCAGCAGCCTGCAACCTGAAG
ATTTTGCAACTTACTACTGTCAACAGAGTAACAGCTGGCCCAACACCTTC
GGCGGAGGGACCAAGGTGGAAATAAAA

130 huT0A-1 VL v1.8 (CDRs DIQMTQSPSSLSASVGDRVTITCRASQTISNNLHWYQQKPGEAPKLLIKSAS
underlined defined by AbM) LAISGVPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQSNSWPNTFGG
GTKVEIK

131 huT0A-1 VL v1.8 GACATCCAGATGACCCAGTCTCCATCCTCCCTGTCTGCATCTGTCGGAG
ACAGAGTCACCATCACTTGCAGGGCCAGCCAAACTATTAGTAACAACCT
GCACTGGTATCAGCAGAAACCAGGCGAGGCCCCTAAGCTCCTGATCAA
GTCTGCTTCCCTGGCCATCTCTGGAGTCCCATCCCGCTTCAGCGGCAG
CGGATCCGGCACAGATTTCACTCTCACCATCAGCAGCCTGCAACCTGAA
GA III! GCAACTTACTACTGTCAACAGAGTAACAGCTGGCCCAACACCTT
CGGCGGAGGGACCAAGGTGGAAATAAAA

132 huT0A-1 VL v2.0 (CDRs EIVMTQSPATLSVSPGERATLSCRASQTISNNLHWYQQKPGQAPRLUYSAS
underlined defined by AbM) LAISGIPARFSGSGSGTEFTLTISSLQSEDFAVYYCQQSNSWPNTFGGGTKV
EIK

133 huT0A-1 VL v2.0 GAAATCGTGATGACACAGTCTCCAGCCACCCTGTCTGTGTCTCCAGGCG
AACGCGCCACCCTGTCCTGCAGGGCCAGCCAAACTATTAGTAACAACCT
GCACTGGTACCAGCAGAAACCTGGCCAGGCTCCCAGGCTCCTGATCTA
TTCTGCTTCCCTGGCCATCTCTGGCATCCCAGCCCGCTTCAGCGGCAG
CGGATCCGGCACAGAGTTCACTCTCACCATCAGCAGCCTGCAGTCCGA
AGATTTTGCTGTGTATTACTGTCAACAGAGTAACAGCTGGCCCAACACCT
TCGGCGGAGGGACCAAGGTGGAAATAAAA

134 Human TrkA, NP 002520 MLRGGRRGQL GWHSWAAGPG SLLAWLILAS AGAAPCPDAC
NP_002520 Lena': 796 CPHGSSGLRC TRDGALDSLH HLPGAENLTE LYIENQQHLQ
June 8, 2009 11:29 Type: P HLELRDLRGL GELRNLTIVK SGLRFVAPDA FHFTPRLSRL
NLSFNALESL
Check: 1056 .. SWKTVQGLSL QELVLSGNPL HCSCALRWLQ RWEEEGLGGV
PEQKLQCHGQ GPLAHMPNAS CGVPTLKVQV PNASVDVGDD
VLLRCQVEGR GLEQAGWILT ELEQSATVMK SGGLPSLGLT
LANVTSDLNR KNVTCWAEND VGRAEVSVQV NVSFPASVQL
HTAVEMHHWC IPFSVDGQPA PSLRWLFNGS VLNETSFIFT EFLEPAANET
VRHGCLRLNQ PTHVNNGNYT LLAANPFGQA SASIMAAFMD
NPFEFNPEDP IPVSFSPVDT
NSTSGDPVEK KDETPFGVSV AVGLAVFACL FLSTLLLVLN KCGRRNKFGI
NRPAVLAPED GLAMSLHFMT LGGSSLSPTE GKGSGLQGHI
IENPQYFSDA CVHHIKRRDI VLKWELGEGA FGKVFLAECH NLLPEQDKML
VAVKALKEAS ESARQDFQRE AELLTMLQHQ HIVRFFGVCT
EGRPLLMVFE YMRHGDLNRF LRSHGPDAKL LAGGEDVAPG

, PLGLGQLLAV ASQVAAGMVY LAGLHFVHRD LATRNCLVGQ
GLVVKIGDFG MSRDIYSTDY YRVGGRTMLP IRWMPPESIL YRKFTTESDV
WSFGVVLWEI FTYGKQPVVYQ LSNTEAIDCI TQGRELERPR ACPPEVYAIM
RGCWQREPQQ RHSIKDVHAR LQALAQAPPV YLDVLG

135 Human TrkB, MSSWIRWHGP AMARLWGFCW LVVGFWRAAF ACPTSCKCSA
NP_001018074 SRIWCSDPSP GIVAFPRLEP NSVDPENITE IFIANQKRLE
IINEDDVEAY
NP_001018074 Length: 822 VGLRNLTIVD SGLKFVAHKA FLKNSNLQHI NFTRNKLTSL
SRKHFRHLDL
December 1, 2007 13:31 SELILVGNPF TCSCDIMWIK TLQEAKSSPD TQDLYCLNES
SKNIPLANLQ
Type: P Check: 9157 IPNCGLPSAN LAAPNLTVEE GKSITLSCSV AGDPVPNMYW
DVGNLVSKHM NETSHTQGSL RITNISSDDS GKQISCVAEN
LVGEDQDSVN LTVHFAPTIT FLESPTSDHH WCIPFTVKGN PKPALQWFYN
GAILNESKYI CTKIHVTNHT EYHGCLQLDN PTHMNNGDYT LIAKNEYGKD
EKQISAHFMG WPGIDDGANP NYPDVIYEDY GTAANDIGDT TNRSNEIPST
DVTDKTGREH LSVYAWVIA SVVGFCLLVM LFLLKLARHS KFGMKGPASV
ISNDDDSASP LHHISNGSNT PSSSEGGPDA VIIGMTKIPV IENPQYFGIT
NSQLKPDTFV QHIKRHNIVL KRELGEGAFG KVFLAECYNL CPEQDKILVA
VKTLKDASDN ARKDFHREAE LLTNLQHEHI VKFYGVCVEG DPLIMVFEYM
KHGDLNKFLR AHGPDAVLMA EGNPPTELTQ SQMLHIAQQI
AAGMVYLASQ HFVHRDLATR NCLVGENLLV KIGDFGMSRD
VYSTDYYRVG GHTMLPIRWM PPESIMYRKF TTESDV\NSLG
VVLWEIFTYG KQPVVYQLSNN EVIECITQGR VLQRPRTCPQ
EVYELMLGCW QREPHMRKNI KGIHTLLQNL AKASPVYLDI LG

136 Chimeric TrkB (d5TrkA) MSSWIRWHGP AMARLWGFCW LVVGFWRAAF
ACPTSCKCSA
SRIWCSDPSP GIVAFPRLEP NSVDPENITE IFIANQKRLE IINEDDVEAY
VGLRNLTIVD SGLKFVAHKA FLKNSNLQHI NFTRNKLTSL SRKHFRHLDL
SELILVGNPF TCSCDIMWIK TLQEAKSSPD TQDLYCLNES SKNIPLANLQ
IPNCGLPSAN LAAPNLIVEE GKSITLSCSV AGDPVPNMYW
DVGNLVSKHM NETSHTQGSL RITNISSDDS GKQISCVAEN
LVGEDQDSVN LTVVNVSFPA SVQLHTAVEM HHWCIPFSVD
GQPAPSLRWL FNGSVLNETS FIFTEFLEPA ANETVRHGCL
RLNQPTHVNN GNYTLLAANP FGQASASIMA AFMGWPGIDD
GANPNYPDVI YEDYGTAAND IGDTTNRSNE IPSTDVTDKT GREHLSVYAV
VVIASVVGFC LLVMLFLLKL ARHSKFGMKG PASVISNDIDD SASPLHHISN
GSNTPSSSEG GPDAVIIGMT KIPVIENPQY FGITNSQLKP DTFVQHIKRH
NIVLKRELGE GAFGKVFLAE CYNLCPEQDK ILVAVKTLKD ASDNARKDFH
REAELLTNLQ HEHIVKFYGV CVEGDPLIMV FEYMKHGDLN KFLRAHGPDA
VLMAEGNPPT ELTQSQMLHI AQQIAAGMVY LASQHFVHRD
LATRNCLVGE NLLVKIGDFG MSRDVYSTDY YRVGGHTMLP
IRWMPPESIM YRKFTTESDV WSLGVVLWEI FTYGKQPWYQ LSNNEVIECI
TQGRVLQRPR TCPQEVYELM LGCWQREPHM RKNIKGIHTL
LQNLAKASPV YLDILG*

137 Chimeric TrkA (d5TrkB) MLRGGRRGQL GWHSWAAGPG SLLAWLI LAS
AGAAPCPDAC
CPHGSSGLRC TRDGALDSLH HLPGAENLTE LYIENQQHLQ
HLELRDLRGL GELRNLTIVK SGLRFVAPDA FHFTPRLSRL NLSFNALESL
SWKTVQGLSL QELVLSGNPL HCSCALRWLQ RWEEEGLGGV
PEQKLQCHGQ GPLAHMPNAS CGVPTLKVQV
PNASVDVGDD VLLRCQVEGR GLEQAGWILT ELEQSATVMK
SGGLPSLGLT LANVTSDLNR KNVTCWAEND VGRAEVSVQV
NVLTVHFAPT ITFLESPTSD HHWCIPFTVK GNPKPALQWF YNGAILNESK
YICTKIHVTN HTEYHGCLQL DNPTHMNNGD YTLIAKNEYG KDEKQISAHF
MDNPFEFNPE DPIPVSFSPV DTNSTSGDPV EKKDETPFGV
SVAVGLAVFA CLFLSTLLLV LNKCGRRNKF GINRPAVLAP EDGLAMSLHF
MTLGGSSLSP TEGKGSGLQG HIIENPQYFS DACVHHIKRR DIVLKWELGE
GAFGKVFLAE CHNLLPEQDK MLVAVKALKE ASESARQDFQ
REAELLTMLQ HQHIVRFFGV CTEGRPLLMV FEYMRHGDLN
RFLRSHGPDA KLLAGGEDVA PGPLGLGQLL AVASQVAAGM
VYLAGLHFVH RDLATRNCLV GQGLVVKIGD FGMSRDIYST
DYYRVGGRTM LPIRWMPPES ILYRKFTTES DVWSFGVVLW
EIFTYGKQPW YQLSNTEAID CITQGRELER PRACPPEVYA
IMRGCWQREP QQRHSIKDVH ARLQALAQAP PVYLDVLG*

138 Chimeric TrkB (d4TrkA) MSSWIRWHGP AMARLWGFCW LVVGFWRAAF ACPTSCKCSA
SRIWCSDPSP GIVAFPRLEP NSVDPENITE IFIANQKRLE IINEDDVEAY
VGLRNLTIVD SGLKFVAHKA FLKNSNLQHI NFTRNKLTSL SRKHFRHLDL
SELILVGNPF TCSCDIMWIK TLQEAKSSPD TQDLYCLNES SKNIPLANLP
NASCGVPTLK VQVPNASVDV GDDVLLRCQV EGRGLEQAGW
ILTELEQSAT VMKSGGLPSL
GLTLANVTSD LNRKNVTCWA ENDVGRAEVS VQVNVHFAPT

HTEYHGCLQL DNPTHMNNGD YTLIAKNEYG KDEKQISAHF
MGWPGIDDGA NPNYPDVIYE DYGTAANDIG DTTNRSNEIP
STDVTDKTGR EHLSVYAVVV IASVVGFCLL VMLFLLKLAR HSKFGMKGPA

ITNSQLKPDT FVQHIKRHNI VLKRELGEGA FGKVFLAECY NLCPEQDKIL
VAVKTLKDAS DNARKDFHRE AELLTNLQHE HIVKFYGVCV EGDPLIMVFE
YMKHGDLNKF LRAHGPDAVL MAEGNPPTEL TQSQMLHIAQ
QIAAGMVYLA SQHFVHRDLA TRNCLVGENL LVKIGDFGMS
RDVYSTDYYR VGGHTMLPIR WMPPESIMYR KFTTESDVWS

PQEVYELMLG CWQREPHMRK NIKGIHTLLQ NLAKASPVYL DILG*

139 Chimeric TrkA (d4TrkB) MLRGGRRGQL GWHSWAAGPG SLLAWLILAS AGAAPCPDAC
CPHGSSGLRC TRDGALDSLH HLPGAENLTE LYIENQQHLQ
HLELRDLRGL GELRNLTIVK SGLRFVAPDA FHFTPRLSRL NLSFNALESL
SWKTVQGLSL QELVLSGNPL HCSCALRWLQ RWEEEGLGGV
PEQKLQCHGQ GPLAHMQIPN CGLPSANLAA PNLTVEEGKS
ITLSCSVAGD PVPNMYVVDVG NLVSKHMNET SHTQGSLRIT
NISSDDSGKQ ISCVAENLVG EDQDSVNLTV SFPASVQLHT
AVEMHHWCIP FSVDGQPAPS LRWLFNGSVL NETSFIFTEF LEPAANETVR
HGCLRLNQPT HVNNGNYTLL AANPFGQASA SIMAAFMDNP
FEFNPEDPIP VSFSPVDTNS
TSGDPVEKKD ETPFGVSVAV GLAVFACLFL STLLLVLNKC GRRNKFGINR
PAVLAPEDGL AMSLHFMTLG GSSLSPTEGK GSGLQGHIIE
NPQYFSDACV HHIKRRDIVL KVVELGEGAFG KVFLAECHNL
LPEQDKMLVA VKALKEASES ARQDFQREAE LLTMLQHQHI
VRFFGVCTEG RPLLMVFEYM RHGDLNRFLR
SHGPDAKLLA GGEDVAPGPL GLGQLLAVAS QVAAGMVYLA
GLHFVHRDLA TRNCLVGQGL VVKIGDFGMS RDIYSTDYYR
VGGRTMLPIR WMPPESILYR KFTTESDVWS FGVVLWEIFT
YGKQPVVYQLS NTEAIDCITQ GRELERPRAC PPEVYAIMRG
CWQREPQQRH S1KDVHARLQ ALAQAPPVYL DVLG*

140 TrkB, Cat (Fells ATGTCGTCCTGGACGAGGTGGCATGGACCCGCCATGGCGCGGCTCTG
domesticus) nucleotide GGGCTTCTGCTGGCTGGTTGTGGGCTTCTGGAGGGCCGCTCTCGCCTG
TCCCACGTCCTGCAAGTGCACCGCCTCTCGGATCTGGTGCAGCGACCC
TTCTCCGGGCATCGTGGCGTTTCCGAGGTTGGAGCCTAATAGTGCAGA
CCCTGAGAACATCACCGAAATTTACATTGCCAATCAGAAAAGGTTGGAA
ATCATCAACGAAGATGATGTCGAAGCTTACGCAGGACTGAAAAATCTGA
CAATTGTGGATTCTGGATTAAAATTTGTGGCTCATAAAGCGTTTCTGAAA
AACAGCAACTTACAGCACATCAA I I I I ACTCGAAATAAACTGACCAGCTT
GTCTAGGAAACATTTTCGTCACCTTGATTTGTCTGAACTGATCCTGGTGG
GCAATCCATTTACATGCTCCTGTGACATTATGTGGATCAAGACTCTTCAG
GAGACTAAATCCAGCCCAGAAACTCAGGATTTGTACTGCCTAAATGAAA
GCAGCAAGAATATTCCCCTGGCAAACCTGCAGATACCCAATTGTGGTTT
GCCATCAGCAAATTTGGCCGCACCTAACCTCACTGTGGARGAGGGAAG
GTCTATCACATTATCTTGCAGTGTCTCAGGCGATCCGGTTCCGAATTTGT
ACTGGGATGTCGGTAATCTGGTTTCCAAGCATATGAATGAAACGAGCCA
CACACAGGGCTCCTTAAGGATAACTAACATTTCATCTGATGACAGTGGA
AAGCAGATCTCCTGTGTGGCAGAAAATCTTGTAGGAGAAGACCAAGATT
CTGTCAACCTCACTGTACATTTTGCTCCAACTATCACATTTCTCGAATCT
CCAACCTCAGACCACCACTGGTGCATTCCATTCACTGTGAAAGGCAACC
CCAAACCAGCTCTTCAGTGGTTCTATAATGGGGCGATACTGAATGAGTC
CAAGTACATCTGTACTAAAATCCATGTTACCAATCACACGGAGTACCATG

, GCTGCCTCCAGCTGGATAATCCTACTCACATGAACAATGGGGACTACAA
GTTAGTAGCCAAGAACGAGTATGGGAAGGATGAGAAACAGATTTCTGCT
CACTTCATGGGCTGGCCTGGAATCGTAGATGGTGCCAACCCAAATTATC
CTGATGTAATTTATGAAGATTATGGGACTGCAGCGAATGACATTGGGGA
CACCACGAACAGAAGTAACGAAATCCCTTCCACAGATGTGGCGGACAAA
AGCGGTCGGGAACATCTTTCGGTCTATGCTGTGGTGGTCATTGCGTCTG

ACACTCCAAGTTTGGCATGAAAGGCCCAGCTTCAGTTATCAGCAATGAT
GATGACTCTGCCAGCCCACTCCACCACATCTCCAATGGGAGTAACACCC
CATCATCTTCAGAGGGCGGCCCCGATGCCGTCATTATTGGAATGACCAA
GATTCCTGTCATTGAAAATCCCCAGTACTTTGGCATCACCAACAGTCAGC
TCAAG CCA GA CA CATTTGTTCAA CACATCAA GC GACATAACATTGTTCTG
AAAAGGGAGCTAGGCGAAGGAGCCTTTGGAAAAGTTTTCCTAGCTGAAT
GCTATAACCTCTGTCCTGAGCAGGACAAGATCTTGGTGGCAGTGAAGAC
GCTGAAGGACGCCAGTGACAACGCCCGCAAGGACTTCCACCGTGAGGC
AGAGCTGCTGACCAACCTCCAGCACGAGCACATTGTCAAGTTCTACGGT
GTCTGTGTGGAGGGCGACCCACTCATCATGGTCTTTGAGTACATGAAGC
ACGGGGATCTCAACAAGTTCCTCAGGGCCCACGGGCCTGACGCTGTGC
TGATGGCCGAAGGCAACCCGCCGACAGAGCTGACGCAGTCCCAGATGC
TGCACATCGCCCAGCAGATAGCAGCGGGCATGGTCTACCTGGCGTCCC
AACACTTTGTGCACCGAGATCTGGCCACCCGGAACTGCCTGGTCGGTG
AGAACCTCCTGGTGAAAATCGGGGACTTCGGGATGTCCCGGGACGTGT
ACAGCACTGACTACTACAGGGTCGGTGGCCACACGATGTTACCCATTCG
CTGGATGCC TCCAGAGAGCATCATGTACAGGAAGTTCACCACAGAAAGT
GATGTCTGGAGCCTGGGAGTCGTGTTGTGGGAGATCTTCACGTACGGC
AAACAGCCCTGGTACCAGCTGTCCAACAACGAGGTGATAGAATGCATCA
CTCAGGGCCGAGTCTTGCAGCGACCTAGAACATGCCCCCAGGAGGTGT
ATGAGTTGATGCTGGGGTGCTGGCAGCGAGAGCCCCACATGAGGAAGA
ACATCAAGGGCATCCACACCCTCCTTCAGAACTTGGCCAAGGCATCTCC
GGTCTACCTGGATATTCTGGGCTAG

141 TrkB, Cat (Fells MSSVVTRWHGPAMARLWGFCWLVVGFWRAALACPTSCKCTASRIWCSDP
domesticus) protein: SPG IVAFPRLEPNSADP EN ITEIYIANQ KRLEI IN
EDDVEAYAGLKN LTIVDSGL

MWIKTLQETKSSP ETQDLYCLN ESSKN IP LANLQ IPNCGLP SAN LAAPN LTV
EEG RSITLSCSVSGDPVPN LYVVDVGN LVSKHMN ETSHTQGSLRITN I SSDD
SGKQISCVAENLVGEDQDSVNLTVHFAPTITFLESPTSDHHWCIPFTVKGNP
KPALQWFYNGAILNESKYICTKIHVTNHTEYHGCLQLDNPTHMNNGDYKLV
AKNEYGKDEKQISAHFMGWPGIVDGANPNYPDVIYEDYGTAANDIGDTTNR
SN EIP STDVADKSGREHLSVYAVVVIASVVGFCLLVMLFLLKLARHSKFGMK
GPASVI SNDDDSASP LH H I SN GSNTPSSSEGGPDAVI IG MTKIPVI EN PQYF G
ITNSQLKPDTFVQHIKRHNIVLKRELGEGAFGKVFLAECYNLCPEQDKILVAV
KTLKDASDNARKDEHREAELLTNLQHEHIVKFYGVCVEGDPLIMVFEYMKH
GDLNKFLRAHGPDAVLMAEGNPPTELTQSQMLHIAQQIAAGMVYLASQHFV
HRDLATRNCLVGENLLVKIGDFGMSRDVYSTDYYRVGGHTMLP IRWMPPE
SIMYRKFTTESDVWSLGVVLWEIFTYGKQPVVYQLSN N EVI EC ITQGRVLQR
PRTCPQEVYELMLGCWQREPHMRKN IKG I HTLLQN LAKAS PVYLDI LG
_

142 TrkB, Dog (Canis familiatis, ATGTCGTCCTGGACGAGGTGGCATGGACCCGCCATGGCGCGGCTCTG
XM_851329) GGGCTTCTGCTGGCTGGTCGTGGGCTTCTGGAGGGCTGCCCTCGCCTG
TCCCACGTCCTGCAAATGCAGCGCCTCTAGGATCTGGTGCAGCGACCC
TTCTCCGGGCATCGTGGCGTITCCGAGGTTGGAGCCCAACAGTGCAGA
CCCTGAGAACATCACCGAAATTTACATTGCCAATCAGAAAAGGTTGGAA
ATCATCAATGAAGATGATGTTGAAGCTTATGCAGGACTGAAGAATCTGAC
GATTGTGGACTCTGGATTAAAATTTGTGGCTCATAAAGCATTTCTGAAAA

GCAATCCATTTACATGTTCCTGTGATATTATGTG GATCAAGACTCTTCAG
GAGACTAAATCCAGCCCAGAAAC TCAGGATTTGTAC TGCCTAAATGAAA
GCAGCAAGAATATTCCCCTGGCAAACCTGCAGATACCCAATTGTGGTTT
GCCATCAGCAAATTTGGCTGCACCTAACCTCACCGTGGAGGAGGGAAA
GTCTATCACATTATCTTGTAGTGTTGCAGGCGATCCAGTTCCGAATTTGT

- 114-.
ACTGGGATGTCGGTAATCTGGTTTCCAAACATATGAATGAAACAAGCCA
CATGCAGGGCTCCTTGAGGATAACTAACATTTCATCTGATGACAGTGGA
AAACAAATCTCCTGTGTGGCAGAAAATCTTGTAGGAGAAGACCAAGATT

CCAACCTCAGACCACCACTGGTGCATTCCATTCACTGTGAAAGGCAACC
CCAAACCAGCGCTTCAGTGGTTCTATAACGGGGCAATATTGAATGAGTC
CAAATACATCTGTACTAAAATCCATGTTACCAATCACACGGAGTACCATG
GCTGCCTCCAGCTGGATAATCCCACTCACATGAACAATGGGGACTACAA
GTTAGTAGCCAAGAATGAGTATGGGAAAGATGAGAAACAGATTTCTGCT
CACTTCATGGGCTGGCCTGGAATTGATGATGGTGCCAACCCAAATTATC
CCGACGTAATTTATGAAGATTACGGGACTGCAGCAAATGACATTGGGGA
CACCACAAACAGAAGTAACGAAATCCCTTCTACAGATGTTGCTGACAAAA
GCGGTCGGGAACATCTTTCGGTCTATGCTGTGGTGGTAATTGCATCTGT
GGTGGGATTITGTCTGCTGGTGATGCTGTTTCTGCTGAAGTTGGCAAGA
CACTCCAAGTTTGGCATGAAAGGCCCAGCTTCAGTTATCAGCAATGATG
ATGACTCTGCCAGCCCCCTCCACCACATCTCCAATGGGAGTAACACCCC
ATCATCTTCAGAGGGCGGCCCCGATGCCGTCATCATTGGAATGACCAAG
ATCCCTGTCATTGAAAATCCCCAGTACTTTGGCATCACCAACAGTCAGCT
CAAGCCAGACACATTTGTTCAGCACATCAAGAGACATAACATTGTTCTGA
AAAGGGAGCTAGGCGAAGGAGCCTTTGGAAAAGTTTTCCTAGCTGAATG
CTATAACCTCTGTCCTGAGCAGGACAAGATCTTGGTGGCAGTGAAGACA
CTGAAGGATGCCAGTGACAACGCACGCAAGGACTTTCACCGCGAGGCT
GAGCTGCTGACCAACCTCCAGCACGAGCACATCGTCAAGTTCTATGGTG
TCTGCGTGGAGGGTGACCCGCTCATCATGGTCTTTGAGTACATGAAGCA
CGGGGACCTCAACAAGTTCCTCAGGGCCCATGGGCCTGATGCTGTGCT
GATGGCCGAAGGCAACCCGCCGACGGAGCTCACCCAGTCCCAGATGCT
GCACATTGCCCAGCAGATAGCAGCAGGAATGGTCTACCTGGCGTCCCA
GCACTTTGTGCACCGAGATCTGGCCACCCGCAACTGCCTGGTTGGCGA
GAACCTCCTGGTGAAAATCGGGGACTTCGGGATGTCCCGGGACGTGTA
CAGCACCGACTACTACAGGGTCGGTGGCCACACAATGCTGCCCATTCG
CTGGATGCCTCCAGAGAGCATCATGTACAGGAAGTTCACCACAGAAAGT
GATGTCTGGAGCCTGGGAGTCGTGTTATGGGAGATCTTCACGTACGGC
AAACAGCCCTGGTACCAGCTGTCCAACAACGAGGTGATAGAATGCATCA
CGCAGGGCCGAGTCTTGCAGCGACCTAGAACGTGCCCCCAGGAGGTCT
ATGAGTTGATGCTGGGGTGCTGGCAGCGGGAGCCCCATATGAGGAAAA
ACATCAAGGGTATCCACACCCTCCTTCAGAACTTGGCCAAGGCATCTCC
AGTCTACCTGGATATTCTAGGCTAG

143 TrkB, Dog (Canis familiaris, MSSWTRWHGPAMARLWGFCWLVVGFWRAALACPTSCKCSASRIWCSDP
XM_851329), amino acid SPGIVAFPRLEPNSADPENITEIYIANQKRLEIINEDDVEAYAGLKNLTIVDSGL
sequence:
KFVAHKAFLKNSNLQHINFTRNKLTSLSRKHFRHLDLSELILVGNPFTCSCDI
MWIKTLQETKSSPETQDLYCLNESSKNIPLANLQIPNCGLPSANLAAPNLIV
EEGKSITLSCSVAGDPVPNLYWDVGNLVSKHMNETSHMQGSLRITNISSDD
SGKQISCVAENLVGEDQDSVNLTVHFAPTITFLESPTSDHHWCIPFTVKGNP
KPALQWFYNGAILNESKYICTKIHVTNHTEYHGCLOLDNPTHMNNGDYKLV
AKNEYGKDEKQISAHFMGWPGIDDGANPNYPDVIYEDYGTAANDIGDTTNR
SNEIPSTDVADKSGREHLSVYAVVVIASVVGFCLLVMLFLLKLARHSKFGMK
GPASVISNDDDSASPLHHISNGSNTPSSSEGGPDAVIIGMTKIPVIENPQYFG
ITNSQLKPDTFVQHIKRHNIVLKRELGEGAFGKVFLAECYNLCPEQDKILVAV
KTLKDASDNARKDFHREAELLTNLQHEHIVKFYGVCVEGDPLIMVFEYMKH
GDLNKFLRAHGPDAVLMAEGNPPTELTQSQMLHIAQQIAAGMVYLASQHFV
HRDLATRNCLVGENLLVKIGDFGMSRDVYSTDYYRVGGHTMLPIRWMPPE
SIMYRKFTTESDVWSLGVVLWEIFTYGKQPVVYQLSNNEVIECITQGRVLQR
PRTCPQEVYELMLGCWQREPHMRKNIKGIHTLLQNLAKASPVYLDILG

144 Cat TrkB For Primer GGATCCGCCG CCACCATGTC GTCCTGGACG AGGTGGCATG G

145 Cat TrkB Rev Primer GCGGCCGCCT AGCCCAGAAT ATCCAGGTAG ACCGGAGAT

146 Dog TrkB For Primer GGATCCGCCG CCACCATGTC GTCCTGGACG AGGTGGCATG G

147 Dog TrkB Rev Primer GCGGCCGCCT AGCCTAGAAT ATCCAGGTAG ACTGGAG
Table 1 Sequence list. Dashes ("-") indicate gaps introduced to align the various allotype and isotype domains with Kabat numbering. Alignments were adjusted so that boundaries (spaces) between blocks of 10 residues always fall between residue numbering positions ending in "0" and residue numbering positions ending in "1".
Where a domain does not begin with a residue whose numbering position ends with "1", phonypadding gaps were inserted to adjust the alignment. For example, for (Seq ID 1) the Kabat domain begins at position 114. Three gap characters were inserted to put the alignment in the correct frame (starting with position 111). These three gap residues (corresponding to Kabat positions 111-113) are properly part of the VH domain and would generally be occupied by real amino acids belonging to the VH in a full antibody sequence. Similar padding gaps may be found at the beginning and end of some other sequences.
The sequences herein described for 29D7 (SEQ ID NOs: 51 and 53, and other sequences that comprise those sequences) are those of the antibody TAM-163.
All examples herein described as using antibody 29D7 used the antibody TAM-163.
EXAMPLES
Numbering of residues Wild type amino acid residues are numbered using a Kabat-compatible numbering system, as illustrated in Figure 2. As used herein, the mutations listed in the tables refer to the domain, followed by the residue position according to Kabat-compatible numbering (see for example SEQ ID NOs: 1 (CH1), SEQ ID NO:9 (CO, SEQ ID NO:13 (CH2), and SEQ ID NO:18 (CH3). The identity of wild type residues is noted in IUPAC single letter code before the residue position (e.g. CH1-S188).
The identity of mutant residues is noted after the residue position (CH1-188E).
Where relevant, both wild-type and mutant residue identity is provided (CH1-S188E).
In an alternative notation, the native amino acid may be listed first, followed by the chain and position in parentheses, followed by the substituted amino acid, for example, Ser(H188)Glu.
Example 1 Identification of Sites for Generating Heterodimer-Favoring Mutations in Antibodies The binding affinity between a protein and another molecule can often be changed by modifying the atoms in closest spatial proximity in the bound state. In native antibodies, a CH1 and CL domain are bound to each other and the degree of binding can be significantly influenced by pairs of atoms, one atom in each domain, that are in close contact (less than 5.0A) in the bound state. Changes to the atoms involved in these binding pairs may lead to either increased or decreased binding.
The specific atoms involved can be determined by methods such as NMR
spectroscopy and protein x-ray crystallography. Atoms on one domain in close contact with the other domain may result in attractive or repulsive forces between the two domains depending on the nature of the atom and its local environment. In addition, for a residue (such as Gly) in a first domain having atoms within 12A of a second domain, close contacts with the second domain may occur if the residue is replaced by a different residue (such as Arg) which adopts a different conformation, and such a residue in the first domain is also considered a close contact residue herein. After modification, the new amino acid may be in close contact with residues on the second domain which were not previously in close contact with the first domain, and these residues are also considered close contact residues. For example, a mutation of Ala to Trp on the first domain may cause unfavorable steric interactions with the second domain, which may be relieved by changing a residue on the second domain, where the residue on the second domain was not in close contact prior to the introduction of Trp on the first domain. This principle may be used to design a novel CL (or CH1) domain which does not interact with wild type CH1 (or CL) domain.
A
novel CH1 (or CL) domain which restores interactions with the novel CL (or CH1) domain may then be constructed. A multispecific antibody may use one or more combinations of such novel CL and CH1 domains to ensure correct pairing between each heavy chain and each light chain. Such designs may be based not just on steric interactions, but also electrostatic interactions, or both types of interactions.
Examination of a protein crystal structure using a graphical tool such as Maestro (Maestro, version 9.2, Schrodinger, LLC, New York, NY (2011)) revealed the atoms in direct close contact by measuring inter-atomic distances using the criteria defined above. In the case of the crystal structure in Protein Data Bank (PDB) entry 3QQ9 (DOI: 10.2210/pdb3qq9/pdb), residues in the CL domain that are in close contact with the CH1 domain include, but are not limited to: 116-119, 121, 123-124, 127, 129, 131, 133, 135-138, 160-164, 167, 174-176, 178, 180, 209 (using the numbering scheme described herein; see Figure 2). The residues in CH1 which are similarly in close contact with CL include, but are not limited to: 121-127, 137-140, 143, 145, 169, 172-180, 186, 188, 190, 192, 221 (using the numbering scheme described herein).
Due to uncertainty present in experimental measurements, and differences in the protein surface environment in different crystal forms, examination of other protein structures may show variations in relative atom positions such that examining these structures results in lists of residues substantially similar, but not identical, to the ones given here. For example, in PDB entry 1HZH (Saphire et al., Science 293:1155-59 (2001)) the structure contains two CH1 domains with different local environments, and in one domain Lys221 is within 4.5A of its partner CL
domain, while in the other CH1 domain it is not. Determination of a close contact in one such CH1/CL interface is sufficient to define a residue as a close contact residue.
Multiple computational methods are available for predicting the orientation of modified amino acid side chains, and the relative effect such changes may have on protein/protein interface binding affinity. However, different methods often give different results. In order to compensate for this variability across methods, several methods were employed to identify amino acid changes that might reduce the affinity of CH1/CL binding. The list of potential amino acid residues that would be targeted was then refined based on inspection of structural models.
Example 2 Antibodies with Novel Covalent CH1-CL Disulfide Linkages Bispecific antibodies can contain different heavy and light chains in each Fab arm. For example, if a bispecific antibody has 2 Fab arms, each with a different LC
and HC, producing a bispecific antibody can involve expression of 4 different polypeptides. Due to the possibility of light chains crossing over and pairing with the incorrect heavy chain, even if the heavy chains are modified to favor heterodimerization, cotransfection and expression of 4 different heavy and light chains can still result in undesirable products, as illustrated in Figure 1.
The wild type interface between CH1 and CL is stabilized by a covalent disulfide bond between CH1-C230 and CL-C214. During assembly of a bispecific antibody, if any incorrect HC/LC pairing occurs, the formation of this disulfide bond may help hold the incorrect pairing in place.
The present inventors postulated that if the mispaired antibody arms could not form the native disulfide, it may increase the opportunity for the mispaired chains to dissociate and find a correct partner. To explore this possibility, alternate positions for the disulfide bond were designed. In these designs, incorrectly paired CH1 and CL
domains cannot form a disulfide bond, because the cysteines are too far apart.
When the correct CH1/CL pairing occurs, a disulfide bond can form and help hold the pairing in place.
A custom method was developed to search the interface between CH1 and CL
and evaluate possible disulfide linkages. The method is similar to that of Dani et al.
(Prot. Eng. 16(3): 187-93 (2003)) but performs additional types of analysis to rank the quality of each site.
Pairs of residues, one on each chain, were chosen where the two residues' alpha carbons are within 7.5A (Cal- Ca2 distance), and the two residues' beta carbons are within 6.0A (C131- C112 distance). To remove pairs where the side chains are oriented away from each other, the distance between C131 and CI32 was compared to the distance between C131 and Ca2. If the former distance is larger, the side chains are partially oriented away from each other and thus poor candidates for forming a disulfide; if the former distance was larger by more than 0.5A the pair was dropped.
For surviving pairs, each putative disulfide was modeled structurally in Modeller (Eswar et al., Nuc. Acids Res. 31(13): 3375-80 (2003)) with 9 models built from randomized starting coordinates for the mutated atoms. A control model of the wild type was also constructed in Modeller. All models were superimposed back on the original crystal structure using TM-ALIGN (Zhang and Skolnick, Nuc. Acids Res.
33: 2302-9 (2005)). The models were checked for the introduction of void volumes in the protein core using VOIDOO (Kleywegt and Jones, Acta Cryst, D50: 178-85 (1994)) with probe radii of 1.0A and 1.5A. Small or nonexistent voids were preferred.
The Modeller DOPE Z-score was calculated, with mutant scores as low as for the wild type being preferred. The Ramachandran plots before and after mutation were compared using PROCHECK (Laskowski, Nuc. Acids Res. 29(1): 221-2 (2001)) to detect any degradation in backbone quality caused by the restraints of the covalent disulfide bond. The mutant model ranked best by the Modeller Objective function was compared to the wild type and the largest displacement of any backbone atom in the two residues being mutated was noted, with smaller displacements being preferred.
Finally, the z1, z2, and x3 angles of the mutant cysteines were computed and compared to a distribution of those angles in 4500 high-resolution crystal structures filtered at the 40% sequence identity level. Putative disulfides resulting in models with geometries deviating least from the experimentally observed distributions were preferred.
Several basic designs obtained from this process are listed in Table 2.
Designs Cys2, Cys4, and Cys5 were ranked less favorably either by manual inspection or by the automated procedure described above, and were not pursued further. Design Cys 3 has two variants: Cys3a and Cys3b. In Cys3b, two additional nearby residues were changed to Ile to improve packing around the disulfide (Vi 901 and L1351), because the F174C mutation was predicted to introduce a small cavity in the structure.
Fab ID CHI Residue(s) CL Residue(s) Base CL Sequence Cys1 A139C F116C
Cys2 F174C S162C
Cys3a F174C S176C K Or A
Cys3b F174C, V1901 L1351, S176C K or A
Cys4 V177C Q160C
Cys5 P123C S121C K Of A
Cys6 F122C E123C K Or A
Cys1k A139C T116C A
Table 1 Novel Cys Pairs Forming Inter-Domain Disulfide Bonds. Each row represents a combination of engineered amino acids in the CH1 and CL domains predicted to result in the formation of a nonnatural disulfide bond between the two domains.
Example 3 Expression and purification of Cys altered heavy/light chains Antibody 29D7 was used as a platform to determine whether the three novel heavy/light chain disulfide bridge positions set out in Table 2 (Cys1, Cys3, Cys6) were capable of forming disulphide bonds. 29D7 is a bivalent, monospecific, monoclonal anti-tyrosine kinase receptor B (TrkB) IgG1 antibody (see Qian et al., J.
Neuroscience 26(37):9394-9403 (2006)).
A positive control with native disulphide bridge (between CH-C230 and CL-C214 "29D7") and a negative control with no bridge at all (CH-C230S and CL-C214S:
"29D7 ACys") were also used in the assay design. 29D7 expression cassette genes were partially constructed using de novo gene synthesis and sub-cloned in frame with 29D7 heavy and light chain variable regions in expression vectors using restriction enzyme-ligation based cloning techniques. Light chain genes were cloned in pSMEN3 and heavy chain genes cloned in pSMED2. Suspension HEK293F cells (American Type Culture Collection) were cultured in serum-free FreeStyleTm293 expression medium (Life Technologies). Cells were maintained in a humidified incubator with 7% CO2 at 37 C. Conditioned media were produced from a standard transient HEK293F transfection process. The conditioned media were filtered through a 0.2pm filter prior to purification. Constructs expressed in the 30-50mg/L
range into conditioned media.
Example 4 Purification of 29D7 antibodies expressed in HEK293F cells Filtered conditioned media was loaded onto HiTrapTm Protein A HP column (GE
Life Sciences) equilibrated with PBS-CMF (137mM NaCI, 2.7mM KCI, 8.1mM
Na2HPO4, 2.7mM KH2PO4, pH 7.2). The resin was washed with 10 column volumes of PBS-CMF pH 7.2 before the antibody was eluted with 0-100% linear gradient of protein A Elution Buffer (20mM citric acid, 150mM NaCI, pH 2.5). Peak fractions were neutralized to pH 7.0 with 2M Tris-HCI pH 8.0 and pooled. The material was loaded onto HiLoad TM 16/60 SuperdexTm200 preparative size-exclusion column (GE Life Sciences) equilibrated in PBS-CMF pH 7.2. Peak fractions were pooled, concentrated using 30kDa spin filters (Am icon) and 0.2pm-filtered.
Analytical SEC was performed using SuperdexTm200 10/300 GL column (GE
Life Sciences) connected to Agilent 1100 Series HPLC system. Under non-reducing conditions, SDS-PAGE analysis (Figure 3) revealed that the negative control, Ab - 121 -29D7 Cys Neg, driven by SDS denaturation, collapses into 100kDa heavy and 25kDa light chain components due to the lack of a heavy/light chain disulphide bridge. The positive control ("29D7") with native disulphide bridge exhibits a single band migrating between the 98kDa and 188kDa markers, presumably representing intact 150kDa IgG1 molecule with heavy and light chains bound by a disulphide bridge. The four novel cysteine constructs described in Table 2 (29D7 Cys1, Cys 3a, Cys3b, and Cys6) behave in a similar fashion as the positive control, implying formation of a disulphide bridge at the positions set out in Table 2.
Example 5 Mass spectrometry of antibodies having altered disulfide linkages To determine the effect of mutations introduced for novel covalent CH1-CL
disulfide linkages, intact mass analysis of the various 29D7 constructs was carried out. Purified forms of antibody 29D7 containing the disulfide modifications listed in Table 2 as well as the positive and negative controls were deglycosylated in the presence of PNGaseF, followed by LC/MS as follows. Antibody was incubated with Lys-C (Wako Chemicals USA, Inc) at a protein:enzyme ratio of 400:1 and incubated at 37 C for 20 mins. The digestion reaction was quenched with addition of 0.1%

formic acid in water. The digested sample was analyzed by LC/MS analysis on an Aglient 1100 capillary HPLC coupled with Water Xevo G2 Q-TOF mass spectrometer.
The analytes were loaded onto a Zorbax Poroshell 300SB C3 column (1.0 mm X 75 mm, maintained at 80 C) with 0.1% formic acid, and eluted using a gradient of 98% buffer B (0.1% formic acid in acetonitrile) at a flow rate of 65 pl/min over 4 mins.
Mass spectrometric detection was carried out in positive, sensitivity mode with capillary voltage set at 3.3 kV. Data analysis were performed with MaxEnt 1 function in MassLynx.
For the 29D7 ACys, the base peak was assigned to monomeric light chain, which has a theoretical mass of 23190 Da (Figure 4A). A secondary peak was assigned to a heavy chain dimer with a clipped lysine (theoretical mass 98086 Da).
This result was consistent with initial design to disrupt formation of disulfide bond between CH1 and CL. The result for construct Cys1 is shown in Figure 4B. The base peak corresponds to intact IgG with lysine clipping (theoretical weight 144438 Da). A

=

partially intact IgG with two heavy chains and only one light chain (theoretical weight 121292 Da) was also observed. A similar result was obtained for construct Cys3a where base peak represents intact IgG with lysine clipping (theoretical weight Da) in addition to the detection of a partially intact IgG with two heavy chains and only one light chain (theoretical weight 121199 Da) (Figure 4C).
Two additional constructs showing majority as intact antibody with only residual partially intact IgG with two heavy chains and only one light chain are shown in Figure 4D & E.
Example 6 DSC Analysis of Antibodies Having Altered Disulfide Linkages Thermal stability of antibodies was measured using Differential Scanning Calorimetry (DSC). The 29D7 disulphide variant antibodies described in Table 2 were diluted in the same buffer (137 mM NaCI, 2.7 mM KCI, 8.1 mM Na2HPO4, and 1.47 mM KH2PO4, pH 7.2) to a concentration of 0.3 mg/mL. Samples and buffers (400 pL) were transferred to a 96 well deep well plate and placed in the autosampler of the DSC (Cap-DSC, Microcal/GE Healthcare). Following injection into the instrument, samples were heated from 10 C to 110 C at 100 C/hr. The data were buffer- and baseline corrected prior to fitting to three, non-two-state transitions to determine the melting temperatures (Table 2). All mutants were stable proteins with high Tm values.
Some differences could be observed in the Tm1 and Tm2, assigned to the CH2 and Fab domains respectively.
Construct Tml ( C) Tm2 ( C) Tm3 ( C) 29D7 (WT) 73.7 0.3 77.9 0.1 84.5 0.1 ACys 73.3 0.3 77.2 0.1 84.6 0.1 Cys1 72.5 0.2 75.7 0.1 84.1 0.1 Cys3a 72.8 0.2 77.6 0.1 84.6 0.1 Cys3b 72.9 0.2 77.7 0.1 84.5 0.1 Cys6 71.7 0.2 75.0 0.1 84.2 0.1 Table 2 DSC Analysis of antibodies having altered disulfide linkages. Each melting temperature refers to the melting of a different key interface. In a canonical antibody, Tm3 is the temperature at which the interface of two CH3 domains melts, Tm1 is the temperature at which the interface of the two CH2 domains melts and Tm2 is the temperature at which the interface of the heavy and light chain melts.
In conclusion, the mutants outlined in Table 2 with Fab ID's (Cys1, Cys3a, Cys3b and Cys6) have novel cysteine residues introduced in both the CHI domain of the heavy chain and constant light domain of the kappa light chain and these cysteines are able to form a novel inter-chain disulphide bond which substitutes the intentionally removed native disulphide bond. These designs were then subsequently evaluated for their ability to favor correct light chain pairing in the bispecific antibody context with native disulphide bridge in one Fab arm and novel disulphide bridge in the other Fab arm.
Example 7 Disruptive Mutants Identified Using Rosetta Modelling Multiple modeling methods were used to identify a set of mutations that could be classed as "disrupting mutations", in that the mutation disfavors pairing of the respective CL and CH domain. The mutations were evaluated by differential scanning calorimetry (DSC) (Table 5), (again, using antibody 29D7 as the test antibody). One modeling method involved using the interface energy method implemented in Rosetta (Das et al., Ann. Rev. Biochemistty 77:363-82 (2008)), version 2.3. Several protocols were used with varying degrees of flexibility in the protein. The "RFlex"
protocol allowed side chains near the mutated residue to relax separately in the bound and unbound states. The "ExRFlex" protocol allowed finer extended sampling of amino acid side chain conformations (Rosetta options "-extrachi_cutoff 12", "-ex1 1", "-ex2 1", "-ex3 1", and "-ex4 1"). Amino acid changes predicted to disrupt the inter-chain binding affinity by more than 1 kcal/mol without causing unfavorable intra-chain energy of more than 10 kcal/mol were initially selected ("disrupting mutations"). Some mutations disrupted binding, but also caused an unfavorable intra-chain energy change (for example, +22 kcal/mol for CH1-S188Y; see Table 3).
For CH1-S188Y, inspection suggested that mutating L143 to a smaller residue could relieve this intra-chain strain. Rosetta predicted that combining CH1-S188Y with CH1-L143A would stabilize the CH1 chain (-5.9 kcal/mol) while still disrupting interactions with the light chain. At some positions, such as CH1-A139, all other amino acids were predicted to disrupt the complex (only a subset of results are =

shown in Table 3).
The total number of possible disrupting mutations obtained by modeling was too large for experimental testing, so disrupting mutations were further modeled for the feasibility of designing compensating mutations in the partner chain that could restore binding ("restoring mutations"). For each disrupting mutation, the protocol identified all close contact residues on the opposite chain as described above. For each disrupting mutation, up to several million candidate sequences with restoring mutations (all possible single and double restoring mutant combinations) were modeled in Rosetta. Representative amino acid positions where at least one Rosetta-predicted disrupting mutation was experimentally tested by DSC are shown in Table 3.
Fab ID Protocol Mutation AE Self-chain MG
H/L (kcal/mol) (kcal/mol) H3 RFlex CH1-A139F >100 -6.4 H6 RFlex CH1-A139H >100 -5.2 H10 RFlex CH1-S188W + CH1-L143S 57.7 -5.9 H10b RFlex CH1-S188W > 100 22.2 H10c RFlex CH1-S188W + CH1-L143A 61.2 -5.9 H9 RFlex CH1-S188Y + CH1-L143A 75.2 -5.9 H9b RFlex CH1-S188Y > 100 22.1 H9c RFlex CH1-S188Y + CH1-L143S 73.6 -5.8 L1 ExRFlex CL-E123K 6.5 L3 ExRFlex CL-S131M 7.5 L4 ExRFlex CL-S131H 8.4 L5 ExRFlex CL-S131P 13.4 L8 ExRFlex CL-Li 35W 6.8 L11 ExRFlex CL-S174Q 16.1 L12 ExRFlex CL-S174M 34.6 L14 ExRFlex CL-S176F 46.1 Table 3 Subset of Disrupting Mutations Identified by Rosetta Calculations. In the Fab ID column, "H" indicates mutations in the CH1 domain of Ab 29D7. "L" indicated mutations in the CL domain of Ab 29D7. The numbers (10, 10b, 9 etc) refer to the different mutation constructs. AE is the predicted change in binding energy between CH1 and CL due to the mutations listed, whereas self-chain AAG reflects stabilization or disruption of the chain containing the mutations.
Example 8 Disruptive Mutants Identified Using SCWRL4 Modelling A second modelling method involved using SCWRL4 (Krivov et al., Proteins 77(4): 778-95 (2009)) to predict the positions of side chains of mutated close contact residues in the interface, followed by energy minimization in MacroModel (MacroModel, version 9.9, Schrodinger, LLC, New York, NY (2012)). Two protocols were used with this method, with variations in the SCWRL step. For the "Base"
method only the mutated side chains were adjusted, whereas for the "Repack"
method all side chains were repacked. Results from the "Repack" method were preferred, as they were expected to indicate that a disruption would not be easily alleviated by minor side chain adjustments. The MacroModel step used the OPLS-2005 force field with GB/SA solvation, and allowed free movement of all hydrogen atoms and the mutated residue(s). Other atoms were restrained by a 100 kJ/mol-restraint, but with 0.2A half-width flat bottom on the energy well. For each mutant the SCWRL4 and MacroModel calculations were performed on the bound state and on the unbound individual CH1 and CL domains, and the binding energy was computed as the energy difference between the bound and unbound forms. This method does not directly measure strain on the chain being mutated, so the most promising models were manually inspected for steric clashes, strained bond angles, or other signs of strain and compensating mutations were added where required.
Approximately 40 different variants were modeled and assessed. Promising representative designs identified by this protocol are listed in Table 4 (some mutations were identified by both Rosetta and SCWRL4+MacroModel).
Fab ID CH1/ CL Mutation AE (kJ/mol) H2 CH1-L124R 55.0 H3 CH1-A139F >100 H6 CH1-A139H 94.4 H11 CH1-V190W 39.0 H14 CH1-K221E 61.1 H16 CH1-A139Y + CH1-V190W >100 H17 CH1-V190W + CH1-K221E >100 L1 CL-E123K 49.0 L3 CL-S131M 17.5 L4 CL-S131H >100 L5 CL-S131P 84.1 L8 CL-L135W 22.2 L11 CL-S174Q >100 L12 CL-S174M 72.4 L14 CL-S176F 50.0 Table 4 Subset of Disrupting Mutations Identified by SCWRUMacroModel Calculations, using the "Repack" protocol. AE is the predicted change in binding energy between CH1 and CL due to the mutations listed.
Example 9 Generation of constructs from modelling Production of mutant Ab 29D2 constructs containing each disrupting mutation set (each row in Tables 3-4) was attempted. The CH1 domain is intrinsically disordered, and adopts the normal folded IgG structure only after interaction with CL.
Prior to interaction with the CL, heavy chains are retained in an unfolded state, bound to the chaperone binding immunoglobulin protein (BiP), in the endoplasmic reticulum (Feige et al., MoL Ce// 34(5): 569-79 (2009)). Thus, if the modelled designs fully disrupt the CH1/CL interaction, no material would be isolatable for further characterization. Constructs H2, H3, H6, H10, H11, H16, H17, L1, L3, L4, L5, L8, L11, L12, and L14 expressed sufficiently for purification, indicating no more than partial disruption of CH1/CL binding. Moderately reduced expression (< 4pg/mL
as compared with parental expression of > 15pg/mL) was observed in COS cells for constructs L4 and L8. Expression of 9b, 9c, 10b, and 10c was not attempted.
Example 10 DSC of expressed constructs Based on structural diversity of sites and selection of similar numbers of CH1 and CL sites, a subset of the expressed Ab 29D7 antibody variants of Example 9 was selected for examination by differential scanning calorimetry (DSC) (see Table below). Constructs in PBS (137 mM NaCI, 2.7 mM KCI, 8.1 mM Na2HPO4, and 1.47 mM KH2PO4, pH 7.2) were diluted in the same buffer to a concentration of 0.3 mg/mL. Samples and buffers (400 pL) were transferred to a 96 well deep well plate and placed in the autosampler of the DSC (Cap-DSC, Microcal/GE Healthcare).
Following injection into the instrument, samples were heated from 10 C to 110 C at 100 C/h. The data were buffer- and baseline corrected prior to fitting to three, non-two-state transitions to determine the melting temperatures. Most changes were seen in the 1m2, the transition assigned to the Fab domain. Considering the standard errors shown in Table 5, all constructs in Table 5 were found to have at least slightly reduced thermal stability relative to antibodies lacking the disrupting mutation(s), indicating that mutations in the CH1/CL interface destabilized the antibody.
The constructs with the largest disruptions of CH1 (H2 H10), and of CL (L1, L4), were selected for follow-up work.
In theory, a mispaired antibody with a restoring mutation (especially a 'hole' designed to accommodate a steric 'bump' on the opposite chain) on one chain and a native sequence on the other chain could still form. It was postulated that a disrupting mutation might have to be made on both the CH1 and CL domain in order to disfavor all possible mispair combinations. Thus, L1 and L4 were chosen over H17, despite their smaller degree of disruption in the DSC experiment, because it was preferred to have multiple disrupting options for both CH1 and CL.
ConstructMutations Tm2 ( C) Tm2 ( C AWT) 29D7 (WT) 77.8 0.1 H2 CH1-L124R 69.7 0.1 -8.1 H3 CH1-A139F 77.2 0.1 -0.6 H6 CH1-A139H 76.6 0.1 -1.2 H10 CH1-3188W + CH1-L143S 69.0 0.1 -8.8 H11 CH1-V190W 76.5 0.1 -1.3 H16 CH1-A139Y+ CH1-V190W 77.4 0.1 -0.4 H17 CH1-V190W+CH1-K221E 72.1 0.3 -5.7 L1 CL-E123K 74.5 0.1 -3.3 L4 CL-S131H 75.2 0.1 -2.6 L8 CL-L135W 75.6 0.1 -2.2 L11 CL-S174Q 76.8 0.1 -1.0 L12 CL-S174M 76.6 0.1 -1.2 L14 CL-S176F 76.7 0.1 -1.1 ACys 77.0+0.1 -0.8 Table 5 Differential scanning calorimetry was used to measure melting temperatures of the various disruptive mutants (taken from Tables 3 and 4) to identify clones which had a lower Fab arm Tm compared to wild-type antibody (29D7).
Example 11 Restoring Mutants Restoring mutations were designed using the SCWRL+MacroModel and Rosetta 2 protocols described above, or by using the Rosetta 3 Sequence Tolerance method (Smith and Kortemme, PLoS One 6(7): e20451 (2011)). With SCWRL+MacroModel, residues on the opposite chain from the disrupting mutations were identified by manual inspection and residues potentially increasing the spatial or electrostatic complementarity of the two chains were noted. Double or triple mutant combinations of these residues were enumerated exhaustively with SCWRL+MacroModel, and the binding energies were compared to the wild type sequence to identify amino acid substitutions which significantly reduced the binding energy loss caused by the disruptive mutation(s). On a modern computing cluster, this protocol is capable of evaluating tens of thousands of mutation combinations. For example, > 1000 combinations of restoring mutations were considered for the disruption example of Table 5. For the Rosetta 2 protocol, which evaluated up to millions of mutants, the search for restoring mutations was described above.
The Rosetta 3 Sequence Tolerance method is not designed for finding disrupting mutations, but its design is appropriate for finding restoring mutations. A
model of the disrupting mutation (constructed using one of the other methods) was provided as input, along with a list of nearby residues (determined by manual inspection) where mutations could ameliorate the disrupting effect. The primary output from this protocol is favorability rankings of amino acid types at each of the nearby sites;
specific sequences containing these residues were then modeled using the sequence tolerance software and/or in SCWRL+MacroModel. In addition to the default Boltzmann-weighted sequence ranking implemented in the Sequence Tolerance package, a method was also used where statistics on amino acid frequency at each position were determined by a step function which applies a weight of 1 to the top-scoring 1% of sequences, and 0 weight to the remaining sequences.
For all protocols, a combination of the protocol energy scores and manual inspection of the models was used to select the best designs. Models in which the rotamers of the mutated side chains closely match known rotamers (from the rotomer data distributed with the Maestro software) were preferred during manual inspection.
Multiple restoring designs were often plausible for a particular disrupting mutant.
Representative restoring designs are shown in Table 6. In the case where a native salt bridge is being reversed, such as between CL-E123 and CH1-K221, the assignment of which residue is disrupting and which is restoring may be considered arbitrary.
In summary, for the disrupting sets of mutations chosen after confirmation by DSC (L1, L4, H2, H10) putative restoring mutations were identified using a combination of multiple computational modeling techniques. For L1, a salt bridge in the wild type sequence, reversal of the salt bridge (but with Glu replaced by Asp) was the only design judged to be worth testing. For the other three disrupting designs, multiple plausible restoring designs were identified.
Fab Disrupting Mutations Restoring Mutations Base CL
ID
Sequence R1.1 CL-E123K CH1-K221D K or A
R1.1 CH1-K221D CL-E123K K or A
R4.1 CL-S131H CH1-K145E
R4.2 CL-S131H CH1-L143H + CH1-Q179D + CH1-S186E
R4.3 CL-S131H CH1-K145S + CH1-S186E
H2.1 CH1-L124R CL-F118E + CL-V133G + CL-S176D K or A
H2.2 CH1-L124R CL-F118H + CL-V133N + CL-S176N K or A
H2.3 CH1-L124R CL-F118D + CL-V133M + CL-S176D K or A
H2.4 CH1-L124R CL-Fl 18E + CL-V1 33N K or A
R10.1 CH1-S188W + CH1-L143S CL-V133S + CL-T178S + CL-S131D
R10.2 CH1-S188W+ CH1-L143S CL-V133S + CL-T178G
R10.3 CH1-S188W + CH1-L143S CL-V133Q + CL-T178G + CL-F118H
R10.4 CH1-S188W + CH1-L143S CL-V133M + CL-T178G + CL-S176G
R4.1X, CL-T131H CH1-K145E A
R4.2k CL-T131H CHI-L143H + CH14179D + CH1-S186E A
R4.3X. CL-T131H CH1-K145S + CH1-S186E A
R10.1k CH1-S188W + CH1-L143S CL-V133S + CL-Y178S + CL-T131D A
R10.2k CH1-S188W + CH1-L143S CL-V133S + CL-Y178G A
R10.3k CH1-S188W + CH1-L143S CL-V133Q + CL-Y178G + CL-F118H A
R10.4k CH1-S188W + CH1-L143S CL-V133M + CL-Y178G + CL-S176G A
Table 6 Example Designs with Disrupting and Restoring Mutations. Each row in the table represents a combination of mutant CH1 and mutant CL domains predicted to associate more favorably than the disrupting mutant would associate with its wild type partner. The first two rows illustrate a case of charge swapping for reversing the orientation of a native salt bridge, where the mutation in either column can be considered disrupting in the absence of the other mutation.
Example 12 Designs with Disrupting, Restoring, and Novel Disulfide Mutations An additional set of mutants consisted of incorporating the novel inter-chain disulphide bond designs highlighted in Table 2, with one or more of the designs listed in Table 6, where appropriate based on molecular modeling. In the cases where a novel disulphide was combined into the bispecific design, the native disulphide cysteine residues (CH1-C230 and CL-C214) were both mutated to serine residues to ablate the native disulphide bond.
Most of the combinations appeared complementary, but in some cases residues to be mutated were near each other (increasing risk of unexpected interactions between the mutations) or were identical. For example, the Cys6 design uses the mutation CL-E123C, which means it is not compatible with the CL-E123K

constructs in Table 6. R4.2 and R10.3 both mutate residue CH1-L143. R4.1 and R10.3 do not mutate the same residues, but are structurally adjacent, increasing the risk of unanticipated interactions between them; in the native structure, residue CL-S131 used in R4.1 contacts CH1-L143 used in R10.3. Example compatible designs are shown in the rows of Table 7, with the CH1 and CL columns of each row constituting a paired design.
Fab ID CHI Mutations CL Mutations Base CL
Sequence Ti K221D + F174C + V1901+ C230S E123K + S176C + L1351+ C214S K or A
T2 K145E + F122C + C230S S131H + E123C + C214S
T3 L143H + Q179D + S186E + F174C S131H + L1351+ S176C + C214S
+ Vi901 + C230S
T4 K145S + S186E + A139C + C230S S131H + F116C + C214S
T9 S188W + L143S + F174C + V133S + T178S + S131D + S176C + K

T12 S188W + L143S + F122C + V133M + T178G + S176G + E123C + K

T18 S188W + L143S + F122C + V133S + T178S + S131D + F116C + K
A139C + F174C + C230S E123C + S176C + C214S
K145E + F122C + C230S T131H + E123C + C214S A
T3k L143H + Q179D + S186E + F174C T131H + L1351+ S176C + C214S
+ V1901+ C230S
T4A. K145S + S186E + A139C + C230S T131H + T116C + C214S A
T9X. S188W + L143S + F174C + V133S + Y178S + T131D + S176C + A

T12k S188W + L143S + F122C + V133M + Y178G + S176G + E123C + A

T18A, S188W + L143S + F122C + V133S + Y178S + T131D + T116C + A
A139C + F174C + C230S E123C + S176C + C214S
Table 7 Example Designs with Disrupting, Restoring, and Novel Disulfide Mutations.
Each row in the table represents a combination of mutant CH1 and mutant CL
domains expected to associate more favorably than the disrupting mutant would associate with its wild type partner.
Example 13 Bispecific Antibodies Having Novel Electrostatic Interactions One type of protein interface selectivity design involves electrostatic complementarity, where a positive charge on one side of an interface is paired with a negative charge on the other side of the interface. If an alternate variant of the interface is engineered in which the charges are reversed, selectivity may occur.
In the present example, the paired residues of each domain involved in existing CH-CL domain salt bridges can be reversed between the interacting domains.
One such example is E123K combined with K221D, as in Table 6. In the final bispecific, one binding arm of the antibody has the wild type salt bridge and one has the reversed salt bridge.
The dual-stage design process described above (first find a disrupting mutation or mutations as in Table 5, then find compensating restoring mutations as in Table 6) may also engineer electrostatic selectivity where only one of the two charged residues is present in the native protein, such as in the R4.1 design of Table 6. Here, the native residues are CL-S131 and CH1-K145. The first stage of the process finds a disrupting mutation of the same charge as the Lys, in this case Cy S131H. Then, the second stage mutates the native Lys to the opposite charge as a restoring mutation, CH1-K145E, which creates a favorable electrostatic interaction.
However, it is conceptually possible to design a de novo favorable electrostatic charge interaction where neither native residue is charged, and use this novel charge interaction to drive interface selectivity. In the absence of the second charged residue, introduction of the first charged residue might not be disruptive (unless for other reasons such as steric contact), and thus might not be found by the dual-stage process described above. Therefore, a different process for de novo electrostatic interaction engineering was also used.

A de novo electrostatic interaction might be placed either in the interface core, where there is little or no exposure to bulk solvent, or it might be placed at the boundary where solvent and both protein chains meet. Core regions, including the CH1/CL interface core, are generally hydrophobic and are not an ideal environment for charged side chains. Unless an optimal hydrogen bonding network that fully satisfies the hydrogen bonding potential of both residues can be engineered, the putative charged residues may have an energetic preference for interaction with solvent (where the CH1 and CL domains remain unbound) rather than each other.
On the other hand, if a charge interaction is engineered on the periphery of the interface, charged residues (particularly Lys, Arg, and Glu) may be sufficiently flexible that a charge/charge mispair might allow two similarly charged residues to orient away from each other, with the electrostatic repulsion significantly dampened by intervening solvent. Exposed residues also create a risk of unwanted clearance of a protein if the immune system mounts an anti-drug antibody (ADA) response against the mutated exposed residues. An ideal case for de novo charge interaction design is a conformationally restricted pocket which does not allow significant side chain flexibility, but which is also sufficiently polar such that the interacting charged residues may be stabilized by additional polar interactions with nearby residues or water molecules.
Such a region on the CH1/CL interface was identified, and efforts were focused there. The CH1/CL interface includes two pockets of water molecules which are in close contact with both domains and which are largely shielded from bulk solvent. In PDB entry 3Q09, these water molecules include those labeled as residues CH-292, CH-319, CH-498, CH-504, CH-544, CL-254, CL-279, CL-359, and CL-490 (Figure 16).
These waters contact protein side chains including CH-L124, CH-L143, CH-K145, CH' 0179, CH-S186, CH-S188, CL-S131, CL-V133, CL-S162, CL-S176, CL-T178, and CL-T180. Most of these residues are polar, but only CH-K145 is charged. The SCWRL/MacroModel method was used to evaluate all possible double mutants of these residues involving one mutation to CH1 and one mutation on CL, and where the native residues were mutated to all possible combinations of Arg, Asp, Glu, and Lys.

This procedure engineered both residues of a novel favorable charge interaction in a single design stage. Results from the protocol without full protein repacking were preferred, to favor designs that may be readily accommodated by the native side chain rotamers without requiring significant adjustments. Additional residues in the pocket, such as CH-F174, CH-V177, CL-F118, and CL-Q128 were noted for reference but were not part of the initial design scan.
Inspection of the results showed that mutations at positions CH-L124 and CL-S176 were promising. The wild type orientations of these residues are shown in Figure 5A. Modeling (not shown) indicated that a favorable electrostatic interaction could form from the combination CH-L124K paired with CL-S176D, and a reversed orientation of the charge interaction could form with CH-L124E and CL-S176K.
However, some bad steric contacts were evident in each case. Manual inspection of the models suggested that for the former charge pair, mutations CH-V190S and CL-Vi 33S would alleviate the strain and, in addition, CL-S133 might form an additional hydrogen bond with position CH-K124 and/or CL-D176. Similarly, for the reversed orientation of the interaction, the mutations CH-S188G and CL-V133S were added to improve packing contacts. After production and experimental validation of these designs, the x-ray crystal structure of each charge pair design was determined, and the results are shown in panels B and C of Figure 5.
The above procedure identified a number of additional potential charge interactions that might be favorably formed in either orientation (VH with a positive amino acid and VL with a negative amino acid, or the reverse, hence these may be considered 'reversible' charge interactions). Often, one or more mutated side chains made minor bad contacts with, or were prevented from adopted a preferred rotamer by, another nearby side chain. In these cases, the predicted double mutant structure was submitted to the Rosetta sequence tolerance protocol to optimize the other nearby surrounding residues.
The identified favorable charge interaction designs are shown in Table 8. Each row in this table is a design which may be used to modify a single CH1/CL
interface.
However, the first column indicates preferred pairings of designs, where the two CH1/CL interfaces of an antibody are separately engineered to each contain one of the two paired designs (a 'forward' and 'reverse' orientation of the charge interaction).
The preferred pairings may result in overt charge/charge repulsion if either CL
attempts to associate with the incorrect CH1 domain.
Fab ID Pairing CHI CL Secondary Base CI_ (Secondary Mutation Mutation Mutations Sequence Mutations included) Si 1 L124K S176D CL-V133S, CH1- K or A

Si_rev 1 L124E 5176K CL-V133S, CH1- K or A

S3_rev 3 S188K T178D CH1-L143D K
S4a 4a L143K T178D K
S4a _rev 4a L143D T178R K
S4b 4b L143K T178D K
S4b _rev 4b L143D T178K CL-S176M K

S5_rev 5 L143R S131E CH1-S186A K
S3X, 3k S188E Y178K CH1-L143E A
S3_rev2. 32. S188K Y178D CH1-L143D A
S4ak 4ak L143K Y178D A
S4a revk 4a2. L143D Y178R A
S4bk 4bk L143K Y178D A
S4b revk 4b2. L143D Y178K CL-S176M A
S52. 52,, L143E T131R A
S5_revk 52. L143R T131E CH1-S186A A
Table 8 Charge Interaction Designs at the CH1/CL Interface. Each row in the table lists a CH1/CL residue pair predicted to form a favorable electrostatic interaction between the two domains, as described in Example 13. The fourth column lists any residues which may improve the behavior of the protein, which may include optimizing packing around the charged residues, heavy/light pairing selectivity, stability, expression, or other features. The "_rev" designation indicates reversal of the orientation of the charges on the CH and CL chains.
Example 14 X-ray crystal structure of S1 and Suev To confirm that the molecular modeling correctly predicted the formation of favorable electrostatic interactions, the x-ray crystal structures of the Si and St_rev designs from Table 8 were determined (Figure 5, panels B and C). The designs were each expressed as recombinant Fab molecules (using antibody Ab1 for the variable domains) by transient transfection in HEK-293 cells. Fabs were purified from conditioned media by batch binding to Poros Protein A resin followed by elution with 0.1 M Glycine pH 2.5. Eluted Fab was then purified by size exclusion chromatography on a Superdex 200 16/60 column equilibrated with 20 mM Tris pH
7.0, 50 mM NaCI. Si Fab was crystallized in 100 mM HEPES pH 7.5, 10 % PEG
3350, 200 mM proline and crystallized in space group P21212 with unit cell edges of 106.7, 127.1, 84.5 A. Si_rev formed crystals in space group P42212 (cell edges 118.4, 118.4, 84.2 A) under conditions of 100 mM sodium Citrate pH 5.9, 14 %
PEG
6000.
Data were collected at beamline 17-ID at the Advanced Photon Source. Data were processed using Autoproc (Global Phasing Ltd.). The structures were solved by molecular replacement with Phaser (Phenix) using as a search model the structure of wild-type Fab solved previously in complex with its ligand (data not shown).
The structures were refined using buster (Global Phasing Ltd.) and built using coot. The Si designed Fab diffracted to 1.3 A resolution and was refined to an R factor of 16.8% (19.0% Rfree). The Si_rev crystals diffracted to 2.1 A resolution, and the structure was refined to an R factor of 17.8% (21.7% Rfree).
In the Si_rev design, CL-K176 made favorable electrostatic contacts (3.3 A
and 3.5 A, Figure 5C) with both carboxyl oxygens of CH1-E124. In the Si design, CH1-K124 made similar but slightly longer (3.5 A and 3.7 A) contacts with CL-D176.
This experimental result confirms the theoretical design, with both S1 and St_rev having favorable electrostatic interactions between the key designed residues.

Example 15 Mixing of Fab Arm Engineering Designs An additional set of combinations uses one Fab arm of the antibody engineered as in any of the rows from Table 7, and the other Fab arm of the antibody engineered using the Si amino acid substitutions given in Table 8. Table 9 shows the resulting combinations.

Fab Arm 1 Fab Arm 2 Fab ID Fab ID
Combination CH1 CH1 Mutations C1 Mutations (Table C1 Mutations (Table Mutations 7) 8) K221D + F174C + E123K + S176C + L124K S176D
1 Ti Si V1901+ C230S L1351+ C214S V190S V1335 K145E + F122C + 5131H + E123C L124K 5176D
2 12 Si C230S + C2145 V190S V1335 L143H + Q179D +
S131H + L1351+ L124K 5176D
3 S186E + F174C + T3 Si 5176C+ C214S V190S V1335 V1901 + C230S
K145S + S186E + 5131H + F116C L124K S176D
4 14 Si A139C + C230S + C214S V190S V1335 V1335 + T178S +
S188W + L143S + L124K S176D
5131D + S176C 19 Si F174C + C230S V190S V133S
+ C2145 V133M + 1178G
S188W + L143S + L124K S176D
6 + S176G + T12 Si F122C + C230S V190S V133S
E123C+ C214S
V1335 + T1785 +
S188W + L143S +
S131D + F116C L124K S176D
7 F122C+ A139C+ T18 Si + E123C + S176C V190S V133S
F174C + C230S
+ C214S
K145E + F122C + T131H + E123C Tn..

2X. C230S + C214S (Table Si 12) L143H + Q179D + T131H + L1351+ T3X Si 3. S186E + F174C + S176C
+ C2145 (Table V1901+ C230S 13) K145S + S186E + T131H + T116C 142k 4X A139C + C230S + C214S (Table Si 12) S188W + L143S + V133S + Y178S + T9X

5X F174C + C230S T131D + 5176C Si + C214S
5188W + L143S + V133M + Y178G T12X

6X F122C+ C230S + S176G + Si E123C + C214S
S188W + L1435 + V133S + Y178S + T18?.
F122C + A139C + T131D + T116C L124K S176D
7X. Si F174C + C230S + E123C + S176C V190S V133S
+ C214S
Table 9 Combination of Electrostatic Interaction Design Si from Table 8 and All Designs from Table 7.
Example 16 Design of modified Bispecific Antibody Ab1/Ab2 Antibody 1 (Ab1) specific for antigen 1 (AG1) was mutated such that its Fab 5 arm contained mutations in the CH1 and C-Kappa domains as denoted in Table 8 Fab ID Si, introducing a novel electrostatic interaction at the constant domain interface. Antibody 2 (Ab2) specific for antigen 2 (Ab2) was mutated such that its Fab arm contained mutations in the CH1 and C-Kappa domains as denoted in Table 8 Fab ID S1 REV also introducing a novel electrostatic interaction at the domain interface.
Knobs-into-holes mutations were introduced into the CH3 domain interface to bias heavy chain heterodimerization (see Ridgway et al., supra and Merchant et at., supra). In one CH3 domain CH3-Y370 was mutated to C and CH3-T389 was mutated to W creating a steric protuberance (referred to as the "Knob" chain; residue numbering is compatible with Kabat, as opposed to the EU numbering of the original reference). In the opposite CH3 domain CH3-S375 was mutated to C, CH3-T389 to S, CH3-L391 to A and CH3-Y438 to V creating a cavity (referred to as the "Hole"
chain) and therefore steric complementarity between the two different CH3 domains.

C370 and CH3-C375 form an inter-chain disulphide bond to stabilize the heterodimer.
Appropriate controls were generated whereby the Fab heavy/light chain interfaces bore no mutations (wild type interfaces) but heavy chain heterodimerizing mutations were still present. A total of four chains comprising the heavy chain of Ab1, heavy chain of Ab2, light chain of Ab1 and light chain of Ab2 were simultaneously transfected into mammalian cells and the level of correct light chain pairing assessed via BlAcore based stoichiometry analysis, mass spectrometry and heterogeneity assessment by anion exchange chromatography. Biophysical analysis results were compared to a control containing the heavy chain heterodimerizing mutations but no mutations at the interface between the heavy and light chain. The isotype of the antibody was human IgG1 with hinge/CH2 heavy chain mutations (L247A, L248A and G250A) that ablate effector functions. Additional mutation designs, as set out in Table 9, were also experimentally assessed using the Ab1/Ab2 antibody to evaluate for propensity for correct light chain pairing.
Example 17 Expression of Bispecific Antibodies Bispecific antibody genes were constructed using de novo gene synthesis and restriction enzyme-ligation based cloning techniques. Light chain genes were cloned in pSMEN3 and heavy chain genes cloned in pSMED2. Suspension HEK293F cells (American Type Culture Collection) were cultured in serum-free FreeStyle-11'1'293 expression medium (Life Technologies). Cells were maintained in a humidified incubator with 7% CO2 at 37 C. Conditioned media were produced from a standard transient HEK293F transfection process. The conditioned media were filtered through 0.2pm filter prior to purification. Typically the bispecific antibodies expressed in the range of 5-50mg/L into the conditioned medium.
Example 18 Purification of antibodies expressed in HEK293F cells Filtered conditioned media was loaded onto HiTrap TM Protein A HP column (GE
Life Sciences) equilibrated with PBS-CMF (137mM NaCI, 2.7mM KCI, 8.1mM
Na2HPO4, 2.7mM KH2PO4, pH 7.2). The resin was washed with 10 column volumes of PBS-CMF pH 7.2 before the antibody was eluted with 0-100% linear gradient of protein A Elution Buffer (20mM citric acid, 150mM NaCl, pH 2.5). Peak fractions were neutralized to pH 7.0 with 2M Tris-HCI pH 8.0 and pooled. The material was loaded onto HiLoad TM 16/60 SuperdexTm200 preparative size-exclusion column (GE Life Sciences) equilibrated in PBS-CMF pH 7.2. Peak fractions were pooled, concentrated using 30kDa spin filters (Amicon) and 0.2pm-filtered.
Analytical SEC was performed using SuperdexTm200 10/300 GL column (GE
Life Sciences) connected to Agilent 1100 Series HPLC system. Depending on the antibody v-domain combination, the typical % high molecular weight species ranged from 2-20% and no low molecular weight species were observed other than the predominant peak of interest representing the 150kDa bispecific antibody species.
Example 19 Mass spectrometric analysis of bispecific antibody (Ab1/Ab2) To confirm the generation of bispecific antibody, Fab fragments of Ab1 and Ab2 were analyzed by mass spectrometry. The molecular weights of Fab fragment from Ab1 and Ab2 are defined by their unique amino acid sequences, and accurate molecular weight determination provides evidence for the presence of correctly paired antibodies.
Bispecific antibody was incubated with Lys-C (Wako Chemicals USA, Inc) at a protein:enzyme ratio of 400:1 and incubated at 37 C for 20 minutes. The digestion reaction was quenched with addition of 0.1% formic acid in water. The digested sample was analyzed by LC/MS analysis on an Aglient 1100 capillary HPLC
coupled with Water Xevo G2 Q-TOF mass spectrometer. The analytes were loaded onto a Zorbax Poroshell 300SB C3 column (1.0 mm X 75 mm, maintained at 80 C) with 0.1% formic acid, and eluted using a gradient of 15-98% buffer B (0.1% formic acid in acetonitrile) at a flow rate of 65 pl/min over 4 minutes. Mass spectrometric detection was carried out in positive, sensitivity mode with capillary voltage set at 3.3 kV. Data analysis were performed with MaxEnt 1 function in MassLynx.
Fab analysis of bispecific antibody Ab1/Ab2 demonstrated that majority of detected Fab fragments are correctly paired Ab1 and Ab2 as shown in Figure 6.
-- There was a significant reduction of incorrectly paired light chain between Ab1 and Ab2 with the introduction of our novel mutations at the CH1/CL-Kappa interface. Fc analysis suggests that a majority of heavy chains are composed of one heavy chain from Ab 1 and one heavy chain from Ab 2. No heavy chain homodimers were detected (Figure 7).
-- Example 20 Tandem Anion Exchange and Mass Spectrometric Analysis of Bispecific Antibody (Abl/Ab 2) Using an Agilent Infinity 1290 UHLPC (Agilent Technologies) fitted with a Q-STAT (Tosoh Bioscience), approximately 20 to 30pg of Bispecific Ab1/Ab2 protein, purified by protein A and preparative SEC chromatography, was injected at a flow -- rate of 1mL/min onto the column equilibrated in 20mM Tris pH 8.6. The protein was then eluted with 1M NaCI in 20mM Tris pH 8.6 over a 7 minute linear gradient from 0-100%.
Protein was detected by absorption at 280nm. The results of this analysis are shown in Figure 8. The parental antibodies are shown in Figure 8 panel A.
Parental -- Ab1 (Figure 8A(i)) displays an apparent homogenous single peak. The anion exchange chromatogram for Ab2 (Figure 8A(ii)) shows a population of acidic and basic charge species which elute before and after the main peak respectively, representing heterogeneity caused by post-translational modifications within the Ab2 Fab arm that affect a proportion of the antibody preparation. The bispecific -- heterodimer Ab1/Ab2 antibody is shown in Figure 8 panel B. This chromatogram shows the incorporation of the charge heterogeneity from the parental antibody (Ab2) into the bispecific antibody. The plot consists of a main peak (Peak 1) which represents approximately 60% of the protein species. Peak 2, which accounts for 32% of the remaining protein consists of two sub-peaks (Peak 2A and Peak 2B).
The remaining 8% of protein is divided between two minor peaks.
Fractions containing or enriched for material from Peaks 1, 2A and 2B from bispecific Ab1/Ab2 fractionation were collected and processed for Fab arm isolation as described above and analyzed by mass spectrometry (Figure 9). The analysis of the three anion exchange fractions revealed that peak 1 contained only two Fab arms with correct light chain pairing based on the expected MW. Fraction 2A is enriched for incorrectly paired Fab consisting of Ab 1 heavy chain and light chain from Ab 2 in addition to correctly paired ab 2 Fab. Fraction 2B consists of correctly paired Ab 1 Fab and correctly paired Ab 2 Fab but with the latter containing a post-translational modification inherited from the parental antibody.
Example 21 DSC Analysis Bispecific antibody Ab1/Ab2 in PBS (137 mM NaCI, 2.7 mM KCI, 8.1 mM
Na2HPO4, and 1.47 mM KH2PO4, pH 7.2) was diluted in the same buffer to a concentration of 0.3 mg/mL. Additionally, the protein was dialyzed overnight with two buffer changes into His:Sucrose (20 mM His, 8.5% sucrose, 50 mg/L EDTA, pH
6.0) using 10 kDa cut-off Slide-A-Lyzer dialysis cassettes and subsequently diluted to 0.3 mg/mL. Samples and buffers (400 pL) were transferred to a 96 well deep well plate and placed in the autosampler of the DSC (Cap-DSC, Microcal/GE Healthcare).
Following injection into the instrument, samples were heated from 10 C to 110 C at 100 C/h. The data were buffer- and baseline corrected prior to fitting to two, non-two-state transitions to determine the melting temperatures. Graphically, the thermal profiles in PBS and His:Sucrose are broadly similar. This is also reflected in the Tm values obtained (Table 10).
Buffer Tm1 (*C) Tm2 ( C) PBS 71.7 0.1 75.9 0.1 His:Sucrose 72.3 0.2 76.6 0.1 Table 10 Example 22 Stability of Bispecific Antibody at High Concentrations Bispecific antibody Ab1/Ab2 was dialyzed overnight with two buffer changes into His:Sucrose (20 mM His, 8.5% sucrose, 50 mg/L EDTA, pH 6.0) using 10 kDa cut-off Slide-A-Lyzer dialysis cassettes. The protein was transferred to a Vivaspin 500 concentrator, 10 kDa cut-off and spun at 14,000 g. The final concentration reached was 112 mg/mL. The sample was transferred to a plastic SEC vial and 20 pL
mineral oil was overlaid. The sample was stored in the dark at room temperature. For each time point, the sample was placed in an Agilent 1200 and 1 pL was injected onto a TOSOH QC-PAK 300 column, using PBS (137 mM NaCI, 2.7 mM KCl, 8.1 mM
Na2HPO4, and 1.47 mM KH2PO4, pH 7.2) as a running buffer, flow rate 0.5 mL/min, min run. The area under the peak was monitored with each injection. The average recovery was 106 2%. Given the good recovery, the percent aggregate was obtained by comparing the area under the curve of the monomer peak with that of the aggregate peak. After 14 weeks at room temperature, only 2.4% aggregation was 15 observed (Table 11).
Time point Day Day Day Week Week Week Week Week Week % aggregation 0 0.7 0.1 0.8 1 1.2 1.4 1.4 2.4 Table 11 Example 23 BlAcore Analysis of Antibody 1 (Ab1) x Antibody 2 (Ab2) vs Cytokine 1 x Cytokine 2 Fab arm mutations used from pair 1 from Table 8 (Fab ID S1 and SLIRev) Using a BlAcore Surface Plasmon Resonance biosensor (T200 model; GE
Healthcare) an analysis of binding stoichiometry was conducted. The binding analysis took place using a running buffer consisting of phosphate buffered saline containing 300mM NaCI, 3.4mM EDTA and 0.01% Tween-20. 12,000 RU of an anti-human antibody (GE Healthcare) was immobilized via amine coupling chemistry to a CM5 carboxymethylated dextran chip (GE Healthcare) surface following manufacturer's instructions as supplied with the human antibody capture kit (part BR-1008-39, GE
Healthcare). The anti-human antibody was amine coupled to both the reference and test flow cell. In order to measure binding stoichiometry, 100-200 RU of purified putative bispecific antibody was captured on the test flow cell at a flow rate of lOuL/min for 30-60s at a concentration of 1-10nM. Cytokine 1 was subsequently flowed over both flow cells at a flow rate of 50uL/min saturating concentration over 100x the KD of the cytokine/antibody interaction, for 60s at which point the reaction had reached steady state. The cytokine and test antibody were stripped from the surface using 3M MgC12 which was exposed to the chip surface at lOuL/min for 40s. The injection port was then washed with running buffer prior to the next cycle.
The process described was then repeated using cytokine 2. Based on the molecular weight of the cytokine (MWC), the molecular weight of the antibody (MWA), the amount of test antibody captured (AB-RU) and the observed binding of cytokine at saturation (Rmax Obs), the observed binding stoichiometry (OBST) was calculated.
The equation for this was:
OBST= [Rmax Obs] / [(MWC/MWA) x AB-RU].
In these studies, the known elements of the equation are the MW of the cytokine and antibody, the RU of antibody captured and the RU of cytokine binding at saturation, with the latter two variables measured experimentally. From that information, observed binding stoichiometry was calculated which infers the %
of captured antibody molecules with correctly formed Fab arms for each respective antigen, since Fab arms with incorrect heavy/light chain pairing will result in no detectable binding of a given Fab arm to a given cytokine and hence a reduction in the pooled binding stoichiometry which represents the entire antibody population.
This latter fact was verified from studies where antibody 1 heavy chain was transfected with antibody 2 light chain and vice versa and binding to each antigen tested by ELISA (data not shown). Data for Rmax_Obs was generated from reference subtracted data and adjusted for baseline drift caused by underlying dissociation of the antibody from the chip surface over time and for non-specific binding. The binding stoichiometries were normalized based on the saturation binding to the parental bivalent positive control antibody.
Saturation binding stoichiometries for cytokine 1 and cytokine 2 of putative bispecific antibodies were compared with bivalent monospecific positive controls and a control with no Fab arm engineering which exhibits all permutations of light chain pairing thus impacting the overall stoichiometry. The data (Table 12) show that the Fab arm engineered novel electrostatic interactions increase the correct light chain pairing to at least 90% with binding stoichiometries significantly closer to 1:1 compared to the negative control which lacks the Fab arm engineered electrostatic interactions for bias toward correct light chain association.
Clone Cytokine 1 Binding Ratio Cytokine 2 Binding Ratio Ab1 2(1.77*) 0.00 Ab2 0 2(2.15*) Ab1xAb2 v1.0 n=1 0.88 0.97 Ab1xAb2 v1.0 n=2 0.90 1.02 Ab1xAb2 v2.0*** n=1 0.92 0.99 Ab1xAb2 NEGATIVE** 0.50 0.64 Table 12 Binding stoichiometries for antibody 1 x antibody 2 bispecific versus cytokine 1 and cytokine 2 utilising Fab arm mutations from pair 1, Table 8 (Fab ID Si and S1 Rev). *Measured ratio, used to normalize bispecific ratios. ** Negative control with heavy chain heterodimerizing mutations but no Fab arm engineering for correct light chain association. *** Note for v2.0 Ab2 had a different antibody variable region framework compared with v1Ø
Example 24 Antibody 1 (Ab1) x Antibody 2 (Ab2) vs Cytokine 1 x Cytokine 2 ¨
Assessment of multiple additional designs Fab arm mutations used from Table 7 in one Fab arm and with other Fab arm bearing native interface Using a BlAcore Surface Plasmon Resonance biosensor (T200 model; GE
Healthcare) an analysis of binding stoichiometry was conducted. The binding analysis took place using a running buffer consisting of phosphate buffered saline containing 300mM NaCI, 3.4mM EDTA and 0.01% Tween-20. 12,000 RU of an anti-human antibody (GE Healthcare) was immobilized via amine coupling chemistry to a CM5 carboxymethylated dextran chip (GE Healthcare) surface following manufacturer's instructions as supplied with the human antibody capture kit (part BR-1008-39, GE
Healthcare). The anti-human antibody was amine coupled to both the reference and test flow cell. In order to measure binding stoichiometry, 100-200 RU of putative bispecific antibody was captured from unpurified conditioned medium on the test flow cell at a flow rate of lOuL/min for 30-60s at a concentration of 1-10nM.
Cytokine 1 =

was subsequently flowed over both flow cells at a flow rate of 50uL/min saturating concentration over 100x the KD of the cytokine/antibody interaction, for 60s at which point the reaction had reached steady state. The cytokine and test antibody were stripped from the surface using 3M MgC12 which was exposed to the chip surface at 10uL/min for 30-40s. The injection port was then washed with running buffer prior to the next cycle. The process described was then repeated using cytokine 2.
Based on the molecular weight of the cytokine (MWC), the molecular weight of the antibody (MWA), the amount of test antibody captured (AB-RU) and the observed binding of cytokine at saturation (Rmax_Obs), the observed binding stoichiometry (OBST) was calculated. The equation for this was: OBST= [Rmax_Obs] / [(MWC/MWA) x AB-RU].
In these studies, the known elements of the equation are the MW of the cytokine and antibody, the RU of antibody captured and the RU of cytokine binding at saturation, with the latter two variables measured experimentally. From that information, observed binding stoichiometry was calculated which infers the %
of captured antibody molecules with correctly formed Fab arms for each respective antigen, since Fab arms with incorrect heavy/light chain pairing will result in no detectable binding of a given Fab arm to a given cytokine and hence a reduction in the pooled binding stoichiometry which represents the entire antibody population.
Data for Rmax_Obs was generated from reference subtracted data and adjusted for baseline drift caused by underlying dissociation of the antibody from the chip surface over time and for non-specific binding. The binding stoichiometries were normalized based on the saturation binding to the parental bivalent positive control antibody.
Saturation binding stoichiometries for cytokine 1 and cytokine 2 of putative bispecific antibodies were compared with a control with no Fab arm engineering which exhibits all permutations of light chain pairing thus impacting the overall stoichiometry. The data (Table 13) shows that the Fab arm engineered mutations from Table 7 increase the correct light chain pairing compared to the negative control which has native heavy /light chain Fab arm interface.

Cytokine 1 Binding Ratio Cytokine 2 Binding Ratio Clone N=1 N=2 N= 1 N=2 Ab1xAb2 v1.0 Ti 0.68 0.53 0.99 0.90 Ab1xAb2 v1.0 T2 0.71 0.58 1.00 0.94 Ab1xAb2 v1.0 T3 0.59 0.47 0.96 0.89 Ab1xAb2 v1.0 T4 0.78 0.64 0.97 0.89 Ab1xAb2 v1.0 T9 0.65 0.64 0.98 0.88 Ab1xAb2 v1.0 T12 0.83 ND 0.64 ND
Ab1xAb2 v1.0 T18 0.79 ND 0.64 ND
Ab1xAb2 NEGATIVE* 0.36 0.37 0.86 0.85 Table 13 Binding stoichiornetries for antibody 1 x antibody 2 bispecific versus cytokine 1 and cytokine 2 utilising Fab arm mutations from Table 7 (Fab ID T1-T4, T9, T12 and T18). *Negative control with heavy chain heterodimerizing mutations but no Fab arm engineering for correct light chain association Example 25 Antibody 1 (Ab1) x Antibody 2 (Ab2) vs Cytokine 1 x Cytokine 2 -Assessment of multiple additional designs Fab arm mutations used from Table 7 (in one arm) and electrostatic interaction mutation from Table 8 (Fab ID Si in the other arm) as summarized in Table 9 Using a BlAcore Surface Plasmon Resonance biosensor (T200 model; GE
Healthcare) an analysis of binding stoichiometry was conducted. The binding analysis took place using a running buffer consisting of phosphate buffered saline containing 300mM NaCI, 3.4mM EDTA and 0.01% Tween-20. 12,000 RU of an anti-human antibody (GE Healthcare) was immobilized via amine coupling chemistry to a CM5 carboxymethylated dextran chip (GE Healthcare) surface following manufacturer's instructions as supplied with the human antibody capture kit (part BR-1008-39, GE
Healthcare). The anti-human antibody was amine coupled to both the reference and test flow cell. In order to measure binding stoichiometry, 100-200 RU of putative bispecific antibody was captured from unpurified conditioned medium on the test flow cell at a flow rate of lOuL/min for 30-60s at a concentration of 1-10nM.
Cytokine 1 was subsequently flowed over both flow cells at a flow rate of 50uL/min saturating concentration over 100x the KD of the cytokine/antibody interaction, for 60s at which point the reaction had reached steady state. The cytokine and test antibody were stripped from the surface using 3M MgCl2 which was exposed to the chip surface at 10uUmin for 30-40s. The injection port was then washed with running buffer prior to =

the next cycle. The process described was then repeated using cytokine 2.
Based on the molecular weight of the cytokine (MWC), the molecular weight of the antibody (MWA), the amount of test antibody captured (AB-RU) and the observed binding of cytokine at saturation (Rmax_Obs), the observed binding stoichiometry (OBST) was calculated. The equation for this was: OBST= [Rmax_Obs] / [(MWC/MWA) x AB-RU].
In these studies, the known elements of the equation are the MW of the cytokine and antibody, the RU of antibody captured and the RU of cytokine binding at saturation, with the latter two variables measured experimentally. From that information, observed binding stoichiometry was calculated which infers the %
of captured antibody molecules with correctly formed Fab arms for each respective antigen, since Fab arms with incorrect heavy/light chain pairing will result in no detectable binding of a given Fab arm to a given cytokine and hence a reduction in the pooled binding stoichiometry which represents the entire antibody population.
Data for Rmax_Obs was generated from reference subtracted data and adjusted for baseline drift caused by underlying dissociation of the antibody from the chip surface over time and for non-specific binding. The binding stoichiometries were normalized based on the saturation binding to the parental bivalent positive control antibody.
Saturation binding stoichiometries for cytokine 1 and cytokine 2 of putative bispecific antibodies were compared with a control with no Fab arm engineering which exhibits all permutations of light chain pairing thus impacting the overall stoichiometry. The data (Table 14) shows that select Fab arm engineered mutations from Table 9 increase the correct light chain pairing compared to the negative control which has native heavy /light chain Fab arm interface.
Cytokine 1 (IL-13) Cytokine 2 (IL-4) Binding Ratio Binding Ratio Clone N=1 N=2 N=3 N=1 N=2 N=3 Ab1xAb2 v1.0 T1* 0.88 0.87 0.84 0.93 0.99 0.92 Ab1xAb2 v1.0 T2* 0.93 0.91 0.83 0.98 1.04 0.97 Abl xAb2 v1.0 T3* 0.92 0.86 0.79 0.97 1.00 0.94 Ab1xAb2 v1.0 T4* 0.96 0.87 0.83 0.99 1.02 0.99 Ab1xAb2 v1.0 T9* 0.85 0.75 0.73 0.98 0.98 0.96 Ab1xAb2 NEGATIVE** 0.47 ND ND 0.81 ND ND

=

Table 14 Binding stoichiometries for select antibody 1 x antibody 2 bispecific versus cytokine 1 (IL-13) and cytokine 2 (IL-4) utilising Fab arm mutations from Table 7 combined with electrostatic interaction pairing 1 from Table 8 as summarized in Table 9. * Select Fab ID's from Table 7 combined with electrostatic interaction Fab ID Si.
**Negative control with heavy chain heterodimerizing mutations but no Fab arm engineering for correct light chain association.
Example 26 Modified Bispecific Antibody C5xAb3 An anti-CCL20 antibody (clone C5) specific for human CCL20 was isolated from a phage library and converted to IgG1 format. Its Fab arm contained mutations in the CH1 and C-Kappa domains as denoted in Table 8 Fab ID S1_Rev, introducing a novel electrostatic interaction at the constant domain interface. An anti-antibody (clone Ab3), specific for human ILI 3, was mutated such that its Fab arm contained mutations in the CH1 and C-Kappa domains as denoted in Table 8 Fab ID
Si also introducing a novel electrostatic interaction at the domain interface.
Two different sets of mutations were introduced into the CH3 domain interface to bias heavy chain heterodimerization, either knobs-into-holes (see Ridgway et al., supra and Merchant et al., supra), termed Method 1 (M1) in the following examples, or the heterodimerization method disclosed in Strop et al., supra, and WO
2011/143545, termed Method 2 (M2). For Ml, the anti-1L13 clone Ab3 heavy chain (with Fab ID
Si), the CH3 domain had the following mutations for heavy chain heterodimerization:

Y370 was mutated to C and CH3-T389 was mutated to W creating a steric protuberance (referred to as the "Knob" chain). In the anti-CC L20 clone C5 heavy chain (with Fab ID S1_Rev) the CH3 domain had the following mutations for heavy chain heterodimerization: CH3-S375C, CH3-T389S, CH3-L391A and CH3-Y438V
creating a cavity (referred to as the "Hole" chain) and therefore steric complementarity between the two different CH3 domains. The Cys-370 and Cys-375 form an inter-chain disulphide bond to stabilize the heterodimer. In the M2 design the mutations used include D232R, P441R, and K440R on the anti-IL13 Ab3 heavy chains and D'232E, P'441E, L'391E on the anti-CCL20 C5 heavy chains.
Appropriate controls were generated whereby the Fab heavy/light chain interfaces bore no mutations (wild type interfaces) but heavy chain heterodimerizing mutations (method 1 or method 2) were still present. All antibodies were IgG1 isotype with hinge/CH2 =

- 148 -effector function ablating mutations (L247A, L248A and G250A). A total of four chains comprising the heavy chain of Ab3, heavy chain of C5, light chain of Ab3 and light chain of C5 were simultaneously transfected into mammalian cells and the level of correct light chain pairing was assessed via various biophysical analysis techniques compared to a control containing the heavy chain heterodimerizing mutations but no mutations at the interface between the heavy and light chain. Four separate expressions were carried out. The first (termed "Ab3xC5-M1") consists of the Fab arm mutations discussed above (Fab ID Si and S1 Rev from Table 7) in combination with heavy chain heterodimerization method Ml. The second expression is a control for the first (termed "Ab3xC5-M1-NEGATIVE") with no mutations in the Fab arms but with heavy chain heterodimerizing mutations present (method M1), The third expression (termed "Ab3xC5-M2") consists of the Fab arm mutations discussed above (Fab ID Si and S1 Rev from Table 7) in combination with heavy chain heterodimerization method M2. The fourth expression is a control for the third (termed "Ab3xC5-M2-NEGATIVE") with no mutations in the Fab arms but with heavy chain heterodimerizing mutations present (method M2). By comparing the level of correct light chain pairing present in control versus test, the effect of the mutations can be assessed. Bispecific antibody CCL20 X Ab3 was expressed and purified as discussed above in Examples 16 and 17 for Ab1/Ab2.
Example 27 Mass spectrometric analysis of bispecific antibody C5xAb3 Fab generation and LC/MS analysis of dual arm antibody (C5 & Ab3) constructs were carried out using same methodology as described above for Ab1/Ab2. A total of four constructs as described above were analyzed to determine the existence of heavy and light chains based on Fab molecular weight measurement.
Deconvoluted mass spectra of C5xAb3-M1 and C5xAb3-M1-NEGATIVE
constructs are shown in Figure 10, where a significant amount of incorrectly paired Fab (C5 heavy chain with Ab3 Light chain & C5 light chain with Ab3 heavy chain) was detected in construct C5xAb3-M1-NEGATIVE (24.8% of total intensity comes from incorrectly paired Fab). However, the amount of incorrectly paired Fab was reduced =

- 149 -in the C5xAb3-M1 which has the described electrostatic interaction mutations present in its Fab arms. The level of correctly light chain paired IgG rose to approximately 95%.
Data for constructs C5xAb3-M2 and C5xAb3-M2-NEGATIVE (Figure 11) show that the intensities of incorrectly paired Fabs in construct C5xAb3-M2-NEGATIVE
amounts to 28.5%, where mis-paired Fab in construct C5xAb3-M2 is reduced to 4.6%. The significant reduction of incorrect pairing demonstrates the effectiveness of engineered electrostatic interaction mutations.
Example 28 DSC Analysis Proteins as listed in Table 15 below were received in PBS (137 mM NaCl, 2.7 mM KCI, 8.1 mM Na2HPO4, and 1.47 mM KH2PO4, pH 7.2) and diluted in the same buffer to a concentration of 0.3 mg/mL. Samples and buffers (400 pL) were transferred to a 96 well deep well plate and placed in the autosampler of the DSC
(Cap-DSC, Microcal/GE Healthcare). Following injection into the instrument, samples were heated from 10 C to 110 C at 100 C/h. The data were buffer- and baseline corrected prior to fitting to two or three, non-two-state transitions to determine the melting temperatures. Overall these were all stable proteins with high Tm values.
Protein Tm1 ( C) Tm2 ( C) Tm3 ( C) C5 73.6 0.1 75.5 0.1 83.9 0.1 Ab3 70.4 0.1 74.1 1.1 83.9 0.1 C5xAb3-M1 70.5 0.1 71.5 0.1 n/a C5xAb3-M2 70.5 0.3 72.8 0.2 n/a Table 15 Example 29 pH Reversibility Proteins as listed in Table 16 below were received in PBS (137 mM NaCI, 2.7 mM KCI, 8.1 mM Na2HPO4, and 1.47 mM KH2PO4, pH 7.2) and diluted to 1 mg/mL
using the same buffer. To two 20 pL aliquots, 0.8 pL PBS was added. Two further 20 pL aliquots were acidified to -pH 3.5 by adding 0.8 pL of a 10x protein A
elution buffer (200 mM citric acid, 1.5 M NaCI, pH 2.0). After 24 h at 4 C, a further 0.5 pL
PBS was added to those samples that had had PBS added before, while the acidified samples were neutralized by addition of 0.5 pL of a 2 M Tris pH 8.0 buffer.
Samples were loaded onto an Agilent 1200 system and 15 pL injected over a TOSOH QC-PAK

.*

- 150 -300 column, using PBS (137 mM NaCI, 2.7 mM KCI, 8.1 mM Na2HPO4, and 1.47 mM
KH2PO4, pH 7.2) as a running buffer, flow rate 0.5 mL/min, 15 min run. The percent monomer from each injection was recorded and used to calculate the percent aggregate in each sample. No significant increases in aggregation were observed after acidification.
Sample % aggregate neutral % aggregate acidified sample sample C5 0.85 0.35 0.80 0.14 Ab3 0.55 0.07 0.55 0.07 C5xAb3-M1 1.85 0.07 1.75 0.21 C5xAb3-M2 0.85 0.07 1.15 0.07 Table 16 Example 30 Forced Aggregation Proteins as listed in Table 17 below were received in PBS (137 mM NaCl, 2.7 mM KCI, 8.1 mM Na2HPO4, and 1.47 mM KH2PO4, pH 7.2) and diluted to 1 mg/mL
using the same buffer. Aliquots (20 pL) were placed in a 96 well plate, overlaid with 40 pL mineral oil and incubated at 40 C, 43.9 C, 50 C, 54 C, 60.1 C and 64 C
in a gradient PCR block for 24 h. Following this, aliquots were loaded onto an Agilent 1200 system and 15 pL were injected over a TOSOH QC-PAK 300 column, using PBS (137 mM NaCI, 2.7 mM KCI, 8.1 mM Na2HPO4, and 1.47 mM KH2PO4, pH 7.2) as a running buffer, flow rate 0.5 mL/min, 15 min run. The percent monomer from each injection was recorded and used to calculate the percent aggregate in each sample. Where recovery was low due to significant aggregation, the peak area was used to calculate the percent of aggregate. The monovalent bispecifics showed significant aggregation from 54 C, while the bivalent bispecifics were still stable at this temperature. Monovalent C5xAb3-M2 was more aggregation prone than C5xAb3-M1.
Sample Incubation temperature 40 C 43.9 C 50 C 54 C 60.1 C 64 C
C5 0.5 0.6 0.7 0.3 1.8 81.1 Ab3 0.2 0.2 0.3 1.3 59.8 99.7 C5xAb3-M1 1.7 1.7 4.4 7.9 13.7 99.7 C5xAb3-M2 0.3 0.3 1.8 22.3 84.6 100.0 Table 17

- 151 -Example 31 Biacore Analysis CCL20-Clone 5 (C5) x IL13-Clone Ab3 (Ab3) Antigens: human CCL20 x human IL13 Fab arm mutations used from pair 1 from Table 8 (Fab ID S1 and SLIRev) Using a BlAcore Surface Plasmon Resonance biosensor (T200 model; GE
Healthcare) an analysis of binding stoichiometry was conducted. The binding analysis took place using a running buffer consisting of hepes buffered saline (HBS) containing 500mM NaCl and 0.01% surfactant p20. 1500 RU of recombinant protein A (Pierce) was immobilized via amine coupling chemistry to a CM5 carboxymethylated dextran chip. The recombinant protein A was amine coupled to both the reference and test flow cell. In order to measure binding stoichiometry 100-200 RU of putative bispecific antibody was captured on the test flow cell at a flow rate of 10uUmin for 30-60s at a concentration of 1-10nM. Recombinant human CCL20 (Peprotech) was subsequently flowed over both flow cells at a flow rate of 50uL/min saturating concentration over 100x the KD of the CCL20/antibody interaction, for 60s at which point the reaction had reached steady state. The cytokine and test antibody were stripped from the surface using 10mM Glycine-HCL pH 1.7 which was exposed to the chip surface at 10uL/min for 30s. The injection port was then washed with running buffer prior to the next cycle. The process described was then repeated using recombinant human IL13 (R&D Systems). Based on the molecular weight of the cytokine (MWC), the molecular weight of the antibody (MWA), the amount of test antibody captured (AB-RU) and the observed binding of cytokine at saturation (Rmax_Obs), the observed binding stoichiometry (OBST) was calculated. The equation for this was: OBST= [Rmax_Obs] / [(MWC/MWA) x AB-RU].
In these studies, the known elements of the equation are the MW of the cytokine and antibody, the RU of antibody captured and the RU of cytokine binding at saturation, with the latter two variables measured experimentally. From that information, observed binding stoichiometry was calculated which infers the `)/0 of captured antibody molecules with correctly formed Fab arms for each respective antigen, since Fab arms with incorrect heavy/light chain pairing will result in no detectable binding of a given Fab arm to a given cytokine and hence a reduction in

- 152 -the pooled binding stoichiometry which represents the entire antibody population.
Data for Rmax Obs was generated from reference subtracted data and adjusted for baseline drift caused by underlying dissociation of the antibody from the chip surface over time and for non-specific binding. The binding stoichiometries were normalized based on the saturation binding to the parental bivalent positive control antibody.
Saturation binding stoichiometries for CCL20 and IL13 of putative bispecific antibodies were compared with bivalent monospecific positive controls and a control with no Fab arm engineering which exhibits all permutations of light chain pairing thus impacting the overall stoichiometry. The data (Table 18) shows that the Fab arm engineered electrostatic interactions increase the correct light chain pairing to -95%
with binding stoichiometries close to 1:1 for each cytokine/chemokine.
CCL20 Binding Ratio IL13 Binding Ratio Clone N=1 N=2 N=1 N=2 C5 2(2.15*) 2(2.18*) 0.00 0.00 Ab3 0.00 0.00 2(1.93*) 2 (1.93*) C5xAb3-M1 (n=1) 0.91 0.96 0.98 1.03 C5xAb3-M1 (n=2) 0.89 0.95 0.99 1.03 C5xAb3-M1-NEGATIVE 0.76 0.77 0.71 0.72 C5xAb3-M2 (n=1) 0.95 0.99 0.96 1.00 C5xAb3-M2 (n=2) 0.92 0.99 1.02 1.06 C5xAb3-M2-NEGATIVE 0.81 0.83 0.65 0.67 Table 18 Binding stoichiometries for anti-CCL20-clone5 x anti-IL13-cloneAb3 bispecific antibodies. *Measured ratio. Used to normalize bispecific ratios.
** Negative control with heavy chain heterodimerizing mutations but no Fab arm engineering for correct light chain association. M1 - heavy chain heterodimerization method 1.

heavy chain heterodimerization method 2 Example 32 Separation of Bispecific Antibodies Using Hydrophobic Interaction Chromatography Hydrophobic interaction chromatography was used to assess protein heterogeneity following the two-step antibody purification process from conditioned media of constructs C5xAb3-M1, C5xAb3-M1-NEGATIVE, C5xAb3-M2 and C5xAb3-M2-NEGATIVE. Using an Agilent Infinity 1290 UHLPC (Agilent Technologies) fitted with a ProPac HIC-10 (Dionex), approximately 20 to 30pg of protein was injected at a flow rate of 1mL/min onto the column equilibrated in 100mM sodium phosphate and , CA 02891714 2015-05-15

- 153 -1M ammonium sulfate pH 7Ø The protein was then eluted with 100mM sodium phosphate pH7.0 over a 7 minute linear gradient from 0-100%. Protein was detected by absorption at 280nm. The results of this analysis are shown in Figure 12.
The parental C5 and Ab3 antibodies shown in Figure 12 panel A display an apparent single peak. The heterodimerization approach M2 is shown in Figure 12 panel C
and heterodimerization approach M1 is shown in Figure 12 panel B. The chromatograms on the left for both panel B and C show the incorporation of the heavy-chain heterodimerization mutations alone. The chromatograms on the right are of bispecific antibodies that contain both the heavy-chain and light-chain mutations. These results clearly show a reduction in the heterogeneity with the incorporation of both the heavy-chain and light-chain mutations Example 33 Calculation of Accessible Surface Area When introducing non-human residues into antibodies that may be potentially be administered to a human subject, there is a risk that the human immune system will recognize the modified residues as foreign and generate antibodies against the administered agent (an anti-drug antibody or ADA response, which may result in faster clearance, reduced activity of circulating agent, or both).
One method of potentially minimizing these consequences of an ADA
response is to choose mutations that are largely confined to the core of the agent, meaning that they are not on the surface of the agent, and therefore are inaccessible for binding by an ADA. Therefore, one way to rank the preferability of bispecific designs is to measure the accessible surface area (ASA) of the modified residues. All other factors being equal, a bispecific design with a lower ASA value for its mutated residues should have lower ADA risk than a bispecific design with a higher ASA
value. The ASA of the complementary residue sets of Table 6 were measured, based on the molecular models described above, and also of designs Si and Si_rev from Table 8, based on x-ray crystallographic analysis. As described above, the designs in Table 8 (Example 13) were specifically designed into a buried pocket, which minimizes ASA. Results are shown in Table 19, and were calculated using the molecular surface tool in Maestro 9.7 (Schrodinger, LLC, 2014) or Maestro 9.9

- 154 -(Schrodinger, LLC, 2015) on the high resolution setting after removal of solvent and buffer molecules, using the structure of each involved domain as the context ('entry' as the context setting). The probe radius was set to 2.5A. The radius of a solvent molecule is often chosen as 1.4A; 2.5A was used here to account for experimental coordinate error in x-ray structures, side chain motion not apparent in the x-ray structure, and for the difficulty of an ADA accessing an extremely narrow opening in the surface as opposed to a water molecule as assumed in many other ASA
calculations. A close approach of protein backbones, as observed in beta strand interactions, would generally result in 2.5A or longer hydrogen bond contacts and thus this is an approximate size for the smallest hole which can be penetrated by any protein chemical group wider than an unbranched side chain.
The consequences of an ADA response may vary by indication. In certain diseases where the immune system is suppressed, the risk of an ADA response may be lower, making designs with higher ASA more feasible. In certain diseases where the immune system is overactive, the ADA risk may be higher, thus requiring potential use of a bispecific design with a low (e.g., <50A2, <40A2, <30A2, <20A2, <
10 A2) or zero ASA value. Some designs previously reported (such as by Lewis et al.
and in W02014/150973A1) have higher ASA than certain designs (such as Si and St_rev) reported herein, and such previously reported designs may be more susceptible to an undesired ADA response. The various embodiments disclosed in W02014150973 all have engineered residues whose ASA is at least 148 A2, or larger (calculated from PDB entry 4LLY (claim 1 of W02014150973) or 4LLW (claim 7 of W02014150973) as reported by Lewis etal. For the embodiment disclosed in claim of W02014150973 and related claims, the value reported is an underestimate, because the side chain of residue 1 on the light chain variable domain was disordered in the crystal structure. With the claimed arginine present, modeling suggests an increased value of 230.3).
The single exception is the mutation at position CL-135, which in some variants is Phe rather than the larger Tyr of claim 1; however, this residue is buried and excluding it from the ASA calculation entirely still results in a value of 149 A2 when

- 155 -combining the residues disclosed at claim 1 and claim 7 of W02014150973 (see also Table 19 of W02014150973). In addition, for the crystal structure 4LLY
reported by Lewis (which contains the mutations relevant to W02014150973), residue 1 of the light chain, which is mutated to Arg, has a disordered side chain. If this residue is added by modeling it in its most commonly occurring rotamer conformation (using the rotamer library in the Maestro software), the total ASA of the mutated residues involved in W02014150973 increases further to 230 A2. The exposed surface area of the designs reported herein, (most notably Si and Si_rev), is considerably lower when implemented in the context of a bispecific with two different heavy chain sequences and two different light chain sequences.
In designing antibodies that may be useful as therapeutics with favorable properties, the introduction of some surface accessible foreign residues may be a necessary requirement to impart certain functional characteristics (potentially including but not limited to stability). Thus, any minimization of the ASA
footprint of foreign residues introduced during other engineering steps may be seen as advantageous in reducing the total final ASA of foreign residues.
Construct ASA (A2) Si 3.5 Si_rev 14.5 Si (Arm 1), Si_rev (Arm 2) 18.0 S3 rev 0 S4a/b 0 S4a_rev 0 S4b_rev 0 S5 0.2 S5 rev 0 R1¨.1 110 R4.1 17.7 R4.2 24.1 R4.3 2.0 H10.1 0 H10.2 0 H10.3 0 H10.4 0 Table 19 Accessible surface area for various heavyheavy chain/light chain heterodimerization designs computed with a probe radius of 2.5A (see text) and high resolution setting in Maestro. Only protein atoms were included in the calculation;

- 156 -solvent and ion atoms were removed. With the designs herein, combining Si and S1 rev as analyzed in examples 14, 16, and 26-33 results in a bispecific antibody with low surface exposure of 18.0A2.
Example 34 Measurement of pairing fidelity Using hydrophobic interaction chromatography it was not possible to differentiate between correct bispecific molecule and molecules containing mis-paired light chains for this particular bispecific antibody most likely because molecules with mis-paired light chains do not have sufficiently different propensity for interacting with the HIC column resin. However Anion Exchange Chromatography (Figure 8B) was able to separate fully bispecific antibody (Peak 1 from Figure 8B) from bispecific antibody containing mis-paired light chain/heavy chain interactions (Peak 2A
from Figure 8B). Unfortunately, a post-translational modification (Sulfation) of one of the parental antibodies, which carried through to the bispecific molecule (Peak 2B
from Figure 8B), lead to a change in the bispecific anion exchange elution time such that it was not baseline resolved from bispecific molecule containing mis-paired light chain.
Therefore it was not possible to accurately quantify by anion exchange chromatography the % bispecific antibody. Addition of Figure 8B Peak 1 % AUC
(60%) to the sulfated form (Figure 8B Peak 2B), which amounted to 17%
approximates % correct bispecific antibody to be 78% of the protein preparation.
Table 20 below represents mass spec relative quantification of correctly paired Fab arm in the Ab1xAb2 example preparations.
Sample % Correctly paired Fab domains*
Ab1xAb2 v1.0 (Figure 6A) 90.5 Ab1xAb2 NEGATIVE (Figure 6B) 66.8 Ab1xAb2 v1.0 AEX** Peak 1 Fraction (Figure 9C) 99.7 Ab1xAb2 v1.0 AEX* Peak 2A*** Fraction (Figure 6A) 64.3 Ab1xAb2 v1.0 AEX* Peak 2B** Fraction (Figure 6A) 88.3 Table 20 % correct Fab species with correct heavy and light chain pairing based on % peak intensities following LCMS as depicted in Figure 6 and Figure 9. * %
correct Fab species determined as a fraction of total peak intensity summed from all correct and incorrect Fab species detected. **AEX = Anion exchange chromatography

- 157 -fractions. ***Peaks 2A and 2B were not baseline resolved and the species attributable to those peaks will have leaked into their adjacent overlapping peak for LCMS purposes. The pairing fidelity of C5xAb3 was also analyzed; results are shown in Tables 21 and 22.
Sample % Peak of Interest*
C5 98.5 Ab3 96.7 C5xAb3-M1 92.1 C5xAb3-M1-NEGATIVE 59.3 C5xAb3-M2 86.3 C5xAb3-M2-NEGATIVE 56.2 Table 21 A Peak of Interest of C5xAb3 antibodies after fractionation on a HIC
Ethyl column as depicted in Figure 12. * % peak of interest is defined by the % area under curve representing either bivalent, monospecific IgG parental antibodies (C5 and Ab3) or bivalent bispecific IgG species (all other clones in table).
Sample % Correctly Paired Fab Domains*
C5xAb3-M1 95.0 C5xAb3-M1-NEGATIVE 75.2 C5xAb3-M2 96.4 C5xAb3-M2-NEGATIVE 71.5 Table 22 % correct Fab species with correct heavy chain and light chain pairing based on `)/0 peak intensities following LCMS as depicted in Figure 10 and Figure 11.
* % correct Fab species determined as a fraction of total peak intensity summed from all correct and incorrect Fab species detected.
Example 35 Effect of Secondary Mutations in S1 and St_rev As shown in Table 8, design Si consists of the primary mutations CH1-L124K
and CL-S176D with secondary mutations CL-V133S and CH1-Vi90S. Design St_rev consists of primary mutations CHI-L124E and CL-S176K, with secondary mutations CL-V133S and CH1-S188G. The secondary mutations were designed to optimize side chain packing in the interface. To test whether these mutations contribute to fidelity of light chain pairing, variants of C5xAb3 were generated in which none, some, or all of the secondary mutations were omitted. The C5 Fab arm was used to test variations of S1 rev, while the Ab3 Fab arm was used to test variations of Si. The knobs-into-_ holes (Ridgway et al., supra and Merchant et al., supra) method, termed M1, was used to bias heavy chain heterodimerization of each combination. From Ml, the domain of each Ab3 variant made for this example had the following mutations for

- 158 -heavy chain heterodimerization: CH3-Y370 was mutated to C and CH3-T389 was mutated to W ("knob" chain). From M1, the CH3 domain of each C5 variant made for this example had the following mutations for heavy chain heterodimerization:

5375C, CH3-T389S, CH3-L391A, and CH3-Y438V ("hole" chain).The Cys-370 and Cys-375 form an inter-chain disulphide bond to stabilize the heterodimer. Six constructs were tested to deconvolute the role of the mutations. "Ab3 C5-M1-NEGATIVE", which contained only the M1 mutations described above for C5 and Ab3, but none of the Si or S1_rev mutations. Deconvolute-2, was reused as a control and is also referred to for this example as Deconvolute-1. "Ab3 C5-M1", which contained the M1 mutations as described above, 51_rev in the C5 Fab arm, and Si in the Ab3 arm, was reused as a control and is referred to for this example as Deconvolute-2. Deconvolute-3 was identical to Deconvolute-2, except that it did not include any of the "Secondary Mutations" for 51 and S1_rev as listed in Table 8.
Deconvolute-4 was identical to Deconvolute-2, except that it omitted the secondary mutations of the CH1 domain for both Si and 51_rev. Thus, for clarity, the Ab3 Fab arm of Deconvolute-4 contained CH1-L124K, CL-5176D, and CL-V133S but not CH1-V190S. And, for clarity, the C5 Fab arm of Deconvolute-4 contained CH1-L124E, CL-5176K, and CL-V1335 but not CH1-5188G. The parent monospecific constructs "C5"

and "Ab3" (also referred to as Deconvolute-5 and Deconvolute-6 respectively), having neither M1 mutations nor Si or S1 rev mutations (CH sequence 54 and CL
sequence 9), and were tested as controls to establish the behavior of the monospecific variant of each antibody. All six designs were IgG1 with hinge/CH2 effector function ablating mutations (L247A, L248A and G250A). For designs Deconvolute-1 through Deconvolute-4, a total of four chains comprising the heavy chain of Ab3, heavy chain of C5, light chain of Ab3 and light chain of C5 were simultaneously transfected into mammalian cells. The level of correct light chain pairing was assessed via various biophysical analysis techniques, and compared to the Ab3Ab3 C5-M1-NEGATIVE, C5, and Ab3Ab3 controls. Separate expressions were carried out for the constructs described. By comparing the level of correct light chain pairing present in control vs test, the effect of mutations can be assessed. The Abs were expressed and purified

- 159 -as discussed in Examples 16 and 17 for Ab1/Ab2. Expression of constructs Deconvolute-1 through Deconvolute-6 ranged from 9 to 200 mg/L.
Example 36 Mass spectrometric analysis of S1 and Si_rev Deconvolution Fab generation and LC/MS analysis of dual arm antibody constructs Deconvolute-1 through Deconvolute-6 (described in the previous example) were carried out using same methods as described above for Ab1/Ab2. A total of six constructs as described above were analyzed to determine the pairing of heavy and light chains based on Fab molecular weight.
Deconvoluted mass spectra are shown in Figure 17. Panels A and B show that the monospecific control antibodies, C5C5 and Ab3Ab3, present one predominant peak corresponding to the predicted molecular mass of their respective Fab fragments. The negative control bispecific Ab3Ab3 C5-M1-NEGATIVE (Panel C), which lacks the Si and Si_rev mutations, shows all four possible heavy/light chain pairings: two correct and two incorrect (Ab3 heavy + C5 light, and Ab3 light +

heavy). The analysis indicates that approximately 30% of the sample consists of mispaired Fabs. The positive control bispecific Ab3Ab3 C5-M1 (Panel D), which contains the full set of Si and Si_rev mutations, shows no visible evidence of heavy/light mispairs, which have predicted masses of 47397 and 46906. Small readings at these values (not readily visible in the figure) lead to a prediction of 0.5%
mispairing in this sample. In contrast, Deconvolute-3 (Panel E), in which the secondary Si and Si_rev mutations (see Table 8) in the heavy and light chains are removed relative to Deconvolute-2, shows clear evidence of both possible heavy/light mispairs accounting for approximately 18% of the sample. Finally, Deconvolute-(Panel F) shows that removing the secondary Si and Si_rev mutations from the heavy chain (but leaving them in the light chain) also allows mispaired chains to form at detectable levels, and 11% of the sample is estimated to be mispaired. In summary, the full set of mutations for Si and Si_rev provides the highest fidelity, while partial implementations of the Si and Si_rev designs provide detectable (but smaller) improvement over the negative control lacking CH1/CL bispecific engineering mutations.

- 160 -Example 37 Hydrophobic Interaction Chromatography Analysis of S1 and Si_rev Deconvolution Constructs Hydrophobic interaction chromatography was used to assess protein heterogeneity following the two-step antibody purification process from conditioned media for constructs Deconvolute-1 through Deconvolute-6. Using an Agilent Infinity 1290 UHLPC (Agilent Technologies) fitted with a ProPac HIC-10 (Dionex), approximately 20 to 30pg of protein was injected at a flow rate of 1mL/min onto the column equilibrated in 100mM sodium phosphate and 1M ammonium sulfate pH 7Ø
The protein was then eluted with 100mM sodium phosphate pH7.0 over a 7 minute linear gradient from 0-100%. Protein was detected by absorption at 280nm. The results of this analysis are shown in Figure 18. Control monospecific antibodies "C5"
and "Ab3" each show a sharp main peak (Panels E-F). If the two antibodies are assembled into a bispecific where only the CH3 domains are engineered, there are overlapping peaks instead of a single main peak, indicating heterogeneity in the sample due to various combinations of the heavy and light chains (Panel A).
When the Si and Suev mutations are added to the CH1/CL interface to favor only the correct heavy/light pairing, sample heterogeneity is greatly reduced (Panel B). The Si and Si_rev designs contain primary mutations that directly form electrostatic interactions, as well as supporting secondary mutations as shown in Table 8.
If the heavy chain supporting mutations are removed (Panel C) while the light chain supporting mutations are left intact, the level of heterogeneity is similar by HIC, but differences were detectable by mass spec (Example 36). If all supporting mutations are removed (Panel D), heterogeneity is reduced relative to "Ab3 C5-M1-NEGATIVE", but is still more pronounced relative to Deconvolute-2."Ab3 C5-M1".
Taken together with the mass spec data of example 36, these results collectively illustrate that the least amount of heterogeneity is produced when the full Si and Suev designs are combined.
Example 38 Production of Bispecific Antibodies with Mixed Fab Arm Designs Conceptually, bispecific Fabs could be engineered by using different combinations of the CHi/CL engineering designs described herein, as discussed in = CA 02891714 2015-05-15

- 161 -Example 15. To test this hypothesis, bispecific antibodies having Ab3 and C5 Fab arms were produced, wherein the Ab3 Fab arm contained either no bispecific engineering mutations (negative control) or the Si design of Table 8. The C5 arm contained either no bispecific engineering mutations (negative control), the Si_rev design of Table 8 (positive control), or one of the designs Ti, T2, T3, T4, T9 as specified in Table 7. An additional control containing Si_rev mutations in C5, but no Fab arm bispecific mutations in Ab3, was also prepared. These 9 constructs, summarized and named in Table 23, were all produced as IgG1 with M1 knobs-into-holes (Ridgway et al., supra and Merchant et al., supra) mutations in the Ab3 and C5 heavy chains in the same configuration described in Example 35, and effector function ablating mutations in the CH2 of both heavy chains, as previously described.
For each design, four chains comprising the heavy chain of Ab3, heavy chain of C5, light chain of Ab3 and light chain of C5 were simultaneously transfected into mammalian cells. The level of correct light chain pairing was assessed via various biophysical analysis techniques, and compared to a control containing the heavy chain heterodimerizing mutations but no mutations at the interface between the heavy and light chain. Separate expressions were carried out for the constructs just described. By comparing the level of correct light chain pairing present in control versus test, the effect of the mutations can be assessed. The antibodies were expressed and purified as discussed above in Examples 16 and 17 for Ab1/Ab2.
Expression of the constructs in Table 23 ranged from 4 to 73 mg/L.
Construct Ab3 Fab Mutations C5 Fab Mutations Ab3-S1xC5-T1 Si (see Table 8) T1 (see Table 7) Ab3-S1xC5-T2 Si T2 Ab3-S1xC5-T3 Si T3 Ab3-S1xC5-T4 Si T4 Ab3-S1xC5-T9 Si T9 Ab3-S1xC5-S1rev (AKA "Ab3 C5- Si Si_rev (see Table 8) Ml")" from prior examples) Ab3-S1xC5 Si None Ab3xC5-S1rev None Si_rev Ab3xC5 (AKA "Ab3-05-M1- None None NEGATIVE")" from prior examples) Table 23: Mixing and Matching of Fab Arm Bispecific Engineering Mutations.

- 162 -In addition to the mutations listed here, all constructs contained M1 heavy chain heterodimerization mutations, and effector function ablating mutations, as described for Example 35.
Example 39 DSC Analysis of Mixed Fab Arm Designs Proteins described in Example 38 and listed in Table 24 below were received in PBS (137 mM NaCI, 2.7 mM KCI, 8.1 mM Na2HPO4, and 1.47 mM KH2PO4, pH
7.2) and diluted in the same buffer to a concentration of 0.3 mg/mL. Samples and buffers (400 pL) were transferred to a 96 well deep well plate and placed in the autosampler of the DSC (Cap-DSC, Microcal/GE Healthcare). Following injection into the instrument, samples were heated from 10 C to 110 C at 100 C/h. The data were buffer- and baseline corrected prior to fitting to two or three, non-two-state transitions to determine the melting temperatures. Overall these were all stable proteins with high Tm values. Detailed melting temperature profiles of each antibody are provided in Figure 19.
Protein Tml ( C) Tm2 ( C) Tm3 ( C) Ab3-S1xC5-T1 70.7 0.0 76.6 0.1 Ab3-S1xC5-T2 67.9 0.8 70.1 0.1 76.6 0.1 Ab3-S1xC5-T3 67.7 1.4 70.1 0.0 76.7 0.1 Ab3-S1xC5-T4 68.5 1.4 70.4 0.1 76.6 0.1 Ab3-S1xC5-T9 70.7 0.0 75.4 0.1 Ab3-S1xC5-S1rev 70.8 0.0 73.8 0.4 Ab3-S1xC5 71.1 0.0 75.2 0.0 Ab3xC5-S1rev 70.8 0.3 73.9 0.3 Ab3xC5 71.5 0.0 75.3 0.0 Table 24 Melting Temperature of Mixed Fab Arm Designs Example 40 Biacore Binding Stoichiometry Analysis of Mixed Fab Arm Designs Fab arm mutations used from Table 23 (Example 38) Using a BlAcore Surface Plasmon Resonance biosensor (T200 model; GE
Healthcare) an analysis of binding stoichiometry was conducted as described for Example 31. Saturation binding stoichiometries for CCL20 and IL13 of putative bispecific antibodies were compared with bivalent monospecific positive controls and a control with no Fab arm engineering which exhibits all permutations of light chain pairing thus impacting the overall binding stoichiometry. The data (Table 25) shows that all combinations with engineering in both Fab arms achieved binding = CA 02891714 2015-05-15

- 163 -stoichiometries close to 1:1 for each pair of antibody with the target cytokine and chemokine (no more than 10% variation from 1:1 binding), while combinations lacking engineering in one or both arms had stoichiometries of less than 0.7:1 in one arm.
These results indicate that different combinations of the bispecific designs can be used to reduce heavy/light chain mispairing.
IL13 Binding Ratio CCL20 Binding Ratio Clone N=1 N=2 N=3 N=1 N=2 N=3 Ab3-S1xC5-T1 0.95 0.96 0.92 1.05 1.05 1.09 Ab3-S1xC5-T2 0.90 0.89 0.92 1.08 1.06 1.09 Ab3-S1xC5-T3 0.96 0.94 0.95 0.99 1.03 0.99 Ab3-S1xC5-T4 0.95 0.94 0.94 0.99 0.97 1.01 Ab3-S1xC5-T9 0.97 0.97 0.96 1.04 1.03 1.01 Ab3-S1xC5-S1rev 1.02 1.01 0.99 0.93 0.92 0.93 Ab3-S1xC5 0.51 0.51 0.54 1.04 1.00 1.02 Ab3xC5-S1rev 0.99 0.99 1.02 0.65 0.68 0.68 Ab3xC5 0.51 0.53 0.52 0.91 0.92 0.91 Table 25 Binding stoichiometries for anti-CCL20-clone5 x anti-IL13-cloneAb3 bispecific antibodies. Measurements were made in triplicate (N=1, N=2, N=3).
The first six constructs, with different combinations of bispecific engineering mutations in the C5 arm, all show binding stoichiometries within 10% of the expected 1:1 value.
The remaining three constructs lack CHi/CL bispecific mutations in one or both of the Fab arms, resulting in a drop in binding stoichiometry (underlined). The mutations in each clone are described in Table 23 and the text of Example 37.
Example 41 Mass spectrometric analysis of Mixed Fab Arm Designs Fab generation and LC/MS analysis of dual arm antibody constructs with various combinations of designs in each Fab arm (as described in Table 23 and Example 38) were carried out using same methods as described above for Ab1/Ab2.
A total of 9 constructs as described above were analyzed to determine the pairing of heavy and light chains based on Fab molecular weight. Deconvoluted mass spectra are shown in Figure 20. When the Si design in the Ab3 Fab arm was paired with Ti, T2, T3, T4, or T9 (Figure 20, Panels A-E) in the C5 Fab arm, minimal amounts of mispaired Fab were detected: 0.5% for Ti, T2, and T3; 0.4% for T4, and 1.3%
for T9.
Ab3xS1 paired with Si_rev in the C5 Fab arm also gave high fidelity, with only 3.2%
mispaired (Panel F; note that this bispecific design used the same Fab arm mutations as Deconvolute-2 of Example 37 and is thus a second measurement of the

- 164 -effectiveness of this design). However, when either Fab arm was lacking a bispecific-favoring design, larger amounts of mispaired sample were produced: 19% with Ab3-S1 and native C5 (Panel G), 41% with native Ab3 and C5-S1rev (Panel H), and 35%
with native Ab3 in one Fab arm and native C5 in the other (Panel l). Thus, in each case where different combinations were tried with each Fab arm using a different bispecific-favoring design, sample purity improved significantly as measured by mass spectrometry.
Example 42 Hydrophobic Interaction Chromatography Analysis of Mixed Fab Arms Designs Hydrophobic interaction chromatography was used to assess protein heterogeneity following the two-step antibody purification process from conditioned media for constructs with various combinations of designs in each Fab arm (as described in Table 23 and Example 38). Using an Agilent Infinity 1290 UHLPC
(Agilent Technologies) fitted with a TOSOH Butyl column, approximately 20 to 30pg of protein was injected at a flow rate of 1mL/min onto the column equilibrated in 50mM sodium phosphate and 2M ammonium sulfate pH 7.2. The protein was then eluted with 50mM sodium phosphate pH7.2 over a 7 minute linear gradient from 0-100%. Protein was detected by absorption at 280nm. The results of this analysis are shown in Figure 21. Bispecific antibodies with Si in the Ab3 Fab arm and any of Ti, T2, T3, T4, or T9 in the C5 Fab arm displayed high fidelity of heavy/light chain pairing (Panels A-E). A minor amount of mispairing is apparent as a small tail on the left side of the main peak. This tail on the peak is slightly larger for Si on Ab3 paired with Sl_rev on C5 (Panel F, see arrow). These results are consistent with the mass spectrographic analysis of Example 41 and Figure 20. If one Fab arm (Panels G-H) or both Fab arms (Panel I) did not contain a bispecific-favoring design, larger amounts of mispaired Fab were detected, as indicated by the presence of additional peaks. For reference, Panels J-K show the corresponding profile of the monospecific Ab3 and C5 antibodies on which these bispecific designs were based; both show a sharp single peak. Thus, in each case where different combinations were tried with each Fab arm using a different bispecific-favoring design, sample purity improved

- 165 -significantly over designs lacking the Fab engineering designs, as measured by hydrophobic interaction chromatography.
Example 43 Mouse Anti-TrkB TOA-1 Antibody The invention includes a humanized mouse antibody that specifically binds human TrkB.
Anti-TrkB antibodies were prepared in mice using human and mouse TrkB-extracellular domain antigens and standard methods for immunization Hybridoma cell line producing the TOA-1 antibody was produced by fusion of individual B cells with myeloma cells. The murine TOA-1 antibody, also referred to as "29D7," is disclosed -- in US7,750,122, herein incorporated by reference in its entirety.
The TOA-1 anti-TrkB antibody heavy chain and light chain variable regions were cloned using the SMART cDNA synthesis system (Clontech Laboratories Incof Mountain View, California) followed by PCR amplification The cDNA was synthesized from 1 pg total RNA isolated from TOA-1 hybridoma cells, using oligo (dT) and the -- SMART IIA oligo (Clontech Laboratories Inc.) with POWERSCRIPTTm reverse transcriptase (Clontech Laboratories Inc.) The cDNA was then amplified by PCR
using a primer which anneals to the SMART IIA oligo sequence and mouse constant region specific primer (mouse Kappa for the light chain and mouse IgG1 for the heavy chain) with VENT polymerase (New England Biolabs Incof Ipswich, -- Massachusetts) Heavy and light chain PCR products were subcloned into the pED6 expression vector and the nucleic acid sequence was determined This method is advantageous in that no prior knowledge of the DNA sequence is required In addition, the resultant DNA sequence is not altered by use of degenerate PCR
primers The nucleotide sequences of the TOA-1 heavy chain variable region is set forth as nucleotides 58-411 of SEQ ID NO: 104. The amino acid sequences of the TOA-1 heavy chain variable region is set forth as residues 20-137 of SEQ ID
NO:
105. The nucleotide sequences of the TOA-1 light chain variable region is set forth as nucleotides 61-381 of SEQ ID NO: 106. The amino acid sequences of the TOA-1 light -- chain variable region is set forth as residues 20-137 of SEQ ID NO:107.

= CA 02891714 2015-05-15

- 166 -Example 44 Construction of chimeric TOA-1 Antibody To verify that the mouse heavy and light chain variable region sequences were correct, chimeric TOA-1 antibody was constructed To generate chimeric TOA-1 heavy chain, the nucleotide sequences of the TOA-1 heavy chain variable region (nucleotides 58-411 of SEQ ID NO: 104) was ligated to cDNA encoding the human IgG1 constant domain mutated for minimal effector function These mutations change the human IgG1 amino acid sequence at residues 234, 235 and 237 defined by EU
numbering from leucine, leucine and glycine to alanine, alanine and alanine respectively. Chimeric TOA-1 light chain was constructed be joining the nucleotide sequences of TOA-1 light chain variable region (nucleotides 61-381 of SEQ ID
NO:
106) to DNA encoding the human Kappa constant region. The alanine present at residue 1 of the TOA-1 light chain variable region was changed to aspartic acid which is commonly found at this position and this was then fused to the human Kappa constant region to generate chimeric TOA-1 Al D light chain (nucleotide sequence SEQ ID NO: 108 and amino acid sequence SEQ ID NO: 109). DNA encoding both versions of chimeric TOA-1 antibody was transiently transfected into COS-1 cells to generate protein. The resultant conditioned medium containing the TOA-1 antibody was quantitated by total human IgG sandwich ELISA Activity of chimeric TOA-1 antibody was assessed by direct binding ELISA Direct binding assays were performed by coating ELISA plates with either human or mouse TrkB-extracellular domain protein (R and D Systems), adding serially diluted conditioned medium containing chimeric TOA-1 antibody and detecting the bound antibody with goat-anti-human IgG-HRP (Southern Biotech). Chimeric TOA-1 antibody bound human and mouse TrkB with comparable affinity as the mouse TOA-1 antibody (Figures 23 and 24). Changing the alanine to aspartic acid at position 1 of the TOA-1 light chain variable region did not affect binding properties to human or mouse TrkB
(Figures 23 and 24). Chimeric TOA-1 antibody was purified by standard Protein A
purification techniques from conditioned medium that was generated by transiently transfecting COS-1 cells with DNA encoding chimeric TOA-1.
Example 45 Humanization of mouse TOA-1 Antibody

- 167 -The CDRs of the mouse TOA-1 antibody were identified using the AbM
definition, which is based on sequence variability as well as the location of the structural loop regions. A humanized TOA-1 heavy chain variable region was constructed to include the CDRs of mouse TOA-1 grafted onto a human DP-54 framework region and this amino acid sequence is set forth as SEQ ID NO: 51 huT0A-1 VH v1Ø The huT0A-1 VH 0.0 is encoded by the nucleic acid sequence in SEQ ID NO: 110. Additional mutations of the human framework acceptor sequences are made, for example, to restore mouse residues believed to be involved in antigen contacts and/or residues involved in the structural integrity of the antigen-binding site.
A24T, R72V and 1...79A mutations predicted to be important for preserving TrkB
binding properties were introduced to the DP-54 framework and this amino acid sequence is set forth as SEQ ID NO: 111 and is referred to herein as huT0A-1 VH
v1.1. The huT0A-1 VH 0.1 is encoded by the nucleic acid sequence in SEQ ID NO:

112. Additionally, a humanized TOA-1 heavy chain variable region was constructed to include the CDRs of mouse TOA-1 grafted onto the DP-3 human germline acceptor framework selected on the basis that it is substantially similar to the framework regions of mouse TOA-1 heavy chain variable region and this amino acid sequence is set forth in SEQ ID NO: 113 huT0A-1 VH v2.0 The huT0A-1 VH v2.0 is encoded by the nucleic acid sequence in SEQ ID NO: 114. Similarly, the DPK21 human germline acceptor framework was used to construct a CDR grafted version of humanized TOA-1 light chain variable region since this germline framework exhibits high sequence identity to the TOA-1 light chain variable region and this amino acid sequence is set forth in SEQ ID NO: 132 huT0A-1 VL v2Ø The huT0A-1 VL v2.0 is encoded by the nucleic acid sequence in SEQ ID NO: 133. Another humanized TOA-1 light chain variable region was constructed to include the CDRs of mouse TOA-grafted onto a human DPK9 germline acceptor framework region and this amino acid sequence is set forth as SEQ ID NO: 115 huT0A-1 VL v1Ø The huT0A-1 VL v1.0 is encoded by the nucleic acid sequence in SEQ ID NO: 116. Additionally, K42E, and Y49K mutations predicted to be important for preserving TrkB binding properties were introduced to the DPK9 framework containing the TOA-1 variable light region

- 168 -CDRs and this amino acid sequence is set forth as SEQ ID NO: 117 huT0A-1 VL
v1.1. The huT0A-1 VL v1.1 is encoded by the nucleic acid sequence in SEQ ID
NO:
118. The huT0A-1 comprising VH v1.0 and VL v1.4 is referred to interchangeably herein as huT0A-1 and TAM-163. Other variants based on the DPK9 framework were constructed and their corresponding nucleotide and amino acid sequences are represented by the SEQ ID NOS listed in Table 27. DNA encoding all possible versions of humanized TOA-1 antibody was transiently transfected into COS-1 cells to generate protein. The resultant conditioned medium containing the humanized TOA-1 antibody variants were quantitated by total human IgG sandwich ELISA.
TrkB
binding properties were evaluated using a competition ELISA with biotinylated chimeric TOA-1 antibody and by Surface Plasmon Resonance (SPR: Biacore).
Table 27: Sequence ID listing for the huT0A-1 VL variants huT0A-1 VL Variant SEQ ID NO (Amino SEQ ID NO (Nucleotide) Acid) huT0A-1 v1.2 119 120 huT0A-1 v1.3 121 122 huT0A-1 v1.4 53 123 huT0A-1 v1.5 124 125 huT0A-1 v1.6 126 127 huT0A-1 v1.7 128 129 huT0A-1 v1.8 130 131 Example 46 Evaluation of TrkB Binding Properties of huT0A-1 Variants TrkB binding properties were assessed for the huT0A-1 variants using a competition ELISA assay with biotinylated chimeric TOA-1 antibody. For this assay procedure, a 96-well plate was coated with rhTrkB-ECD (R&D #397-TR/CF) at 1 pg/ml, overnight at 4 C. The plate was then blocked with PBS + 0.02% casein for 1 hour at room temperature Biotinylated chimeric TOA-1 at 25 ng/ml in PBS + 0.5%
BSA + 0.02% tween-20 was mixed with varying concentrations of huT0A-1 variants or unlabeled chimeric TOA-1 and incubated at room temperature for 1 hour. The wells were washed four times with PBS + 0.03% tween-20 Streptavidin-HRP
(Southern Biotech catalog #7100-05) diluted 1:10,000 was added and incubated for

- 169 -30 minutes at room temperature. The wells were washed four times with PBS +
0.03% tween-20 and TMB (BioFx) was added. The reaction developed for 5- 10 minutes and was then quenched with 0.18 N H2SO4. The absorbance at 450nm was determined. Results summarized in Table 28 show that humanized TOA-1 VH
version 1.0 and VL version 1.1 completely retained TrkB binding properties relative to the chimeric TOA-1 antibody (Figures 25, 26 and 27). Further characterization was done to determine which mouse framework residues contained within TOA-1 VL version 1.1 are required for binding TrkB Humanized TOA-1 VL version 1.4 contains a single mouse framework residue K49 (Kabat numbering) and this version has comparable activity to TOA-1 VL version 1.1 (Figure 28) Table 28: Summary of Humanized TOA-1 Variants TrkB Binding Properties VH VL IC50 [nM]
mouse mouse 3.6 mouse human v1.0 16.3 mouse human v1.1 13.6 mouse human v2.0 13.6 human v1.0 mouse 3.6 human v1.0 human v1.0 6.2 human v1.0 human v1.1 4.3 human v1.0 human v2.0 5.9 human v1.1 mouse 3.0 human v1.1 human v1.0 13.6 human v1.1 human v1.1 10.1 human v1.1 human v2.0 10.7 human v2.0 mouse -26.3 human v2.0 human v1.0 -126 human v2.0 human v1.1 -87.6 human v2.0 human v2.0 -37.8 Example 47 Kinetic Evaluation of huT0A-1 Variants BIACORE analysis was performed to determine the affinity constants for TOA-1 and the humanized TOA-1 variants to human and mouse TrkB. BIACORE
technology utilizes changes in the refractive index at the surface layer upon binding of the TOA-1 antibody variants to the TrkB protein immobilized on the layer Binding is detected by surface plasmon resonance (SPR) of laser light refracting from the surface. Analysis of the signal kinetics on-rate and off-rate allows the discrimination

- 170 -between non-specific and specific interactions Human and mouse TrkB ectodomain proteins (R&D Systems, #397/TR/CF and #1494-TB/CF) were immobilized at a low density on a CM5 chip (41 and 30 RUs respectively) and then various concentrations of TOA-1 and humanized TOA-1 variants were injected over the surface. The surface was regenerated with 4M MgC12 between injection cycles. Results show that the humanized TOA-1 variants have comparable affinity constants for both human and mouse relative to both the parental mouse TOA-1 antibody as well as the chimeric TOA-1 antibody (Table 29) demonstrating that these humanized variants have fully retained TrkB binding properties.
Table 29: Kinetic affinity constants for humanized TOA-1 variants binding human or mouse TrkB
Human TrkB Mouse TrkB
Antibody ka kd KD (nM) ka kd KD (nM) muT0A-1 2.1 x 3.1 x 10- 14.7 4.2 x 6.9 x 10-16.3 chiT0A-1 2.8 x 3.4 x 10- 12.1 2.6 x 4.4 x 10-17.0 semi-huT0A-1 4.6 x 3.2 x 10- 7.0 6.8 x 7.4 x 10-10.9 huT0A-1 v1.0/1.0 3.4 x 5.1 x 10- 14.9 2.6 x 4.8 x 10-18.4 huT0A-1 v1.0/1.1 3.5 x 4.2 x 10- 12.1 5.4 x 7.4 x 10-13.8 huT0A-1 v1.0/1.4 4.3x 4.1 x10- 9.4 4.4x 5.0 x 10-11.4 (TAM-163) 105 3 105 3 huT0A-1 v1.0/2.0 3.3 x 5.5 x 10- 16.5 5.3 x 1.1 x 10-20.9 Example 48 Agonist activity of anti-TrkB antibodies The ability of humanized anti-TrkB TOA-1 antibodies to activate the TrkB
signaling cascade was assessed using 1) a transcriptional reporter to monitor TrkB
signalling activation and 2) evaluating autophosphorylation of hTrkB and phosphorylation of ERK1/2, AKT, and PLCy1, known mediators of TrkB signaling (reviewed in Friedman et al Exp Cell Res 1999; 253:131-142).
TrkB expressing stable cell-lines Stable cell lines of HEK-293 cells expressing both a CRE-luciferase reporter and rhuTrkB (nm_006180) or rmuTrkB (nrn_001025074) were generated using _

- 171 -standard techniques (Zhang et al, 2007, Neurosignals 15:29-39). Stable cell lines are designated rhuTrkB-CRE and rmuTrkB-CRE.
Transcriptional reporter assay The luciferase reporter assay was performed as followsrhuTrkB-CRE cells were plated at 35,000 cells/100u1/well in growth media (10% FCS ¨ DMEM) in white bottom 96-well platesThe next day, 1Oul/well of murine TOA1, humanized TOA-1 variants or isotype control (mIgG1 or hIgG1) antibody was added as 10X to the assay plates without changing media Luciferase activity was measured 16-18 hrs later using the Steady-Glo Luciferase Assay System (Promega, E2520) according to the manufacturer's protocol. In brief, media was replaced with 100u1/well of lx PBSNext, 100u1/well of Steady-Glo reagent is added. Plates are sealed with TopSeal (PerkinElmer cat# 6005185) and shaken at A Plate Shaker (IKA Works, Inc.) at speed 600 for 5 minutes. The luminescence was measured using VICTOR 3, 1420 Multilabel Counter (Perkin Elmer).
As exemplified in Figure 29 and summarized in Table 30, TOA-1 antibody treatment resulted in a dose-dependent increase in luciferase activity with mouse TOA-1 and all humanized TOA-1 variants, indicating that these antibodies are able to activate the TrkB signalling cascade Table 30: Agonist activity of Anti-TrkB TOA-1 antibodies in CRE-luciferase reporter assay

- 172 -Humanized TOA-1 variant Relative Activity VH VL
mu mu ++++
mu hu v1.0 ++
mu hu v1.1 ++
mu hu v2.0 ++
hu v1.0 mu +++
hu v1.0 hu v1.0 +++
hu v1.0 hu v1.1 +++
hu v1.0 hu v1.2 ++
hu v1.0 hu v1.3 ++
hu v1.0 hu v1.4 +++
hu v1.0 hu v1.5 +++
hu v1.0 hu v1.6 ++
hu v1.0 hu v1.7 +++
hu v1.0 hu v1.8 +++
hu v1.0 hu v2.0 ++
hu v1.1 mu +++
hu v1.1 hu v1.0 ++
hu v1.1 hu v1.1 hu v1.1 hu v2.0 ++
hu v2.0 mu ++
hu v2.0 hu v1.0 hu v2.0 hu v1.1 hu v2.0 hu v2.0 Stimulation of TrkB autophosphorylation and phosphorylation of PLCy1, AKT, and ERK1/2 by TOA-1 antibodies Phosphorlyation analyses was performed to measure activation of proximal markers of TrkB signaling in engineered cell lines that overexpress TrkB
(rhuTrkB-CRE and rmuTrkB-CRE stable cell lines generated as described above) and differentiated Human SH-SY5Y neuroblastoma cells that express human TrkB.

' CA 02891714 2015-05-15

- 173 -Human TrkB expression in these cells was confirmed by Western analysis using standard techniques, as described below, using an anti-TrkB antibody (BD
Transduction Labs Cat# 610102). TrkB expressing cells were treated with TOA-1 antibodies and Western analyses performed to monitor autophosphorylation of hTrkB
(Tyr490) and phosphorylation of ERK1/2 (Thr202/Tyr204), AKT (Ser473), p38 (Thr180/Tyr182) and PLCy1 (Tyr783) as detailed below.
rhuTrkB-CRE or rmuTrkB-CRE cells were plated in 6-well plates at 5x 105 cells/well in 10% FCS ¨ DMEM growth media and cultured until the cells were 85-90% confluent. Cells were washed once with 0.1`)/0 FCS ¨ DMEM (low serum media) and incubated for an additional 4 hours in low serum media. Next, cells were treated with BDNF (R&D #248BD) or TOA-1 antibody at the designated concentration for 60 min Medium was aspirated from the wells and 0.6 ml of 1X loading buffer (Invitrogen, with 1% b-ME) was added per well to lyse the cells. Cell lysates were transferred to Eppendoff tubes, and heated at 100 C for 5 min 25 ul of each sample were resolved on a NuPAGE 4-20% Bis-Tris gradient gel (Invitrogen).
Western analysis was performed as follows: After electrophoresis, size-fractionated proteins were transferred onto nitrocellulose membranes.
Membranes were blocked with 5% milk in T-TBS (0.15% Tween 20 in TBS), incubated with the appropriate primary antibody [anti-P-TrkB: Phosphor-TrkA (Tyr490),Cell signaling (CS) #9141; anti-Phospho-PLCy1 (Tyr783), CS #2821; anti-Phospho AKT (Ser473), CS#9271; anti-P-ERK1/2 (Phospho-P44/P42 (Thr202/Tyr204), CS #9101; anti-Actin, Sigma A2066] in 1% milk T-TBS on a rocking platform at 4 C overnight Membranes were washed 3X in T-TBS, then incubated with the appropriate HRP-conjugated secondary antibody (Cell Signaling #7974) for 2 hours. Next, membranes were washed 4 times in T-TBS and once in TBS. The signals were developed using ECL
kit (GE RPN2106V) and the manufacturer's protocol followed by x-ray film exposure or Gel-Doc (Bio-Rad) to capture the image.
Human neuroblastoma SH-SY5Y cells were plated in 6-well plates at 2x105cells/well and cultured in growth media (DMEM:F12 (1:1) supplemented with 2mM L-glutamine, 15% FBS and pen/strep). Cells were incubated with retinoic acid

- 174 -(10uM) for 3 days to induce differentiation. Then, the cells were cultured in low serum media (growth media with 1% FBS) overnight, and further cultured in 0.1% FBS
medium for 4 hours BDNF (R&D #248BD) or TOA-1 antibody at the designated concentration was added and cells incubated for 15-60 min. Medium was aspirated from the wells and 0.6 ml of 1X loading buffer (Invitrogen, with 1% b-ME) was added per well to lyse the cells. Cell lysates were transferred to Eppendoff tubes, and heated at 100 C for 5 min 20 pl of each sample was resolved on a NuPAGE 4-20%
Bis-Tris gradient gel (Invitrogen) Western analysis was performed as described above.
As shown in Figure 30, treatment with humanized TOA-1 antibody induced dose-dependent auto-phosphorylation of TrkB and phosphorylation of the signaling molecules ERK1/2, AKT, and PLCy1 in cells expressing human TrkB (both overexpressed TrkB, figure 30A, and endogenous TrkB, figure 30C) or mouse TrkB

(Figure 30B).
Also shown is the relative activity of BDNF and mT0A-1 in the phosphorylation assays performed as described above BDNF is a more potent stimulator of the TrkB
signaling cascade as measured by TrkB autophosphorylation and PLCyl phosphorylation than mT0A-1 (Figure 30D).
In summary, both the transcriptional reporter assay and phosphorylation assays demonstrated that TOA-1 antibodies activate the TrkB signalling pathway.
Example 49 Characterization of antibody binding epitopes relative to BDNF
Competition ELISAs were used to evaluate how the binding of anti-TrkB
antibodies to TrkB protein affects the BDNF interaction with TrkB protein.
In one format, a 96-well plate (Costar, cat# 3590) was coated with BDNF
(0.3ug/rd, R&D systems, cat # 248-BD) in PBS at 4C and incubated overnight. The plate was washed with PBS, 0.1% Tween-20, then blocked wells with PBS, 1% BSA, 0.05%
Tween-20 at Room temperature for 1hour. Multiple concentrations of ProA
purified anti-TrkB antibody were pre-incubated with rhTrkB/Fc Chimera (15Ong/ml, R&D
systems, cat# 688-TK) for 30 minutes at room temperature, then the mixtures were added to the plate and incubated for 1hr at room temperature. The plate was washed = CA 02891714 2015-05-15

- 175 -with PBS, 0.1% Tween-20 6 times, peroxidase conjugated goat anti-human IgG
(Fc) antibody (PIERCE, cat# 31413) was added and incubated for 1hr at room temperatue. The wells were washed with PBS 3 times and the substrate TMB
(BioFX
Laboratories, cat # TMBW-0100-01) added for 10 minutes. The reaction was stopped with 0.18N H2SO4The absorbance at 450nm was determined.
In a second format, a 96-well plate was coated with rhTrkB-ECD-His (1pg/m1) in PBS overnight at 4 C The plate was then blocked with PBS + 0.02% casein for hour at room temperature Biotinylated humanized TOA-1 at 25 ng/ml was incubated with varying concentrations of rhBDNF, unlabeled huT0A-1, or an irrelevant human IgG1 antibody and then the mixtures were added to the plate and incubated for hour at room temperature. The plate was washed with PBS + 0.03% tween-20 four times, Streptavidin-HRP (Southern Biotech catalog #7100-05) diluted 1:10,000 was added and incubated for 30 minutes at room temperature. The wells were washed four times with PBS + 0.03% tween-20 and TMB (BioFx) was added. The reaction developed for 5¨ 10 minutes and was then quenched with 0.18 N H2SO4 The absorbance at 450nm was determined As shown in Figure 31, results using both competition ELISA formats indicate that TOA-1 partially competes with BDNF for binding to human TrkB, suggesting that the TOA-1 binding site at least partially overlaps with the BDNF docking site on hTrkB
Example 50 Mapping the TOA-1 binding site on human TrkB
To further delineate the TOA-1 binding site on TrkB, a series of chimeric TrkB-TrkA receptors were generated and evaluated for TOA-1 binding in a cell-based ELISA.
Generation of TrkB-TrkA chimeric receptor expression constructs TrkA-TrkB chimeric receptors were generated and cloned into the mammalian expression vector pcDNA3.1 (lnvitrogen) using standard molecular cloning techniques. The chimeric TrkB(d5TrkA) receptor (Sequence 35) was generated by replacing residues 284-377 (np_001018074 Sequence 34), referred to as domain 5 of TrkB, with the TrIKA domain 5 residues 280-377 (np_002520, Sequence 33).
Similarly, chimeric TrkA(d5TrkB) was generated by replacing residues 283-377 of = CA 02891714 2015-05-15

- 176 -TrkA with residues 281-377 of TrkB (Sequence 36) Chimeric TrkB (d4TrkA) was generated by replacing residues 190-282 of TrkB with residues 187-281 of TrkA
(Sequence 37) Chimeric TrkA(d4TrkB) was generated by replacing residues 187-of TrkA with residues 190-282 of TrkB (Sequence 38).
Cell-based ELISAs to evaluate the binding of humanized TOA-1 antibodies to TrkB-TrkA chimeric receptors were performed as follows:
Cell transfection:
Human embryonic kidney 293 cells (ATCC) cells are plated at 4.5x10^6 cells per 10cm2 tissue culture plate and cultured overnight at 37C. The next day cells are transfected with Chimeric TrkA-B expression plasmids using LF2000 reagent (lnvitrogen, Cat # 11668-019) at a 3:1 ratio of reagent to plasmid DNA using the manufacturers protocol. Cells are harvested 48hrs after transfection using Trypsin, washed once with phosphate buffered saline (PBS), then suspended in growth media without serum at 2x10e6 cells/ml.
Cell-based ELISA assay Anti-TrkB or control antibodies at 1 pg/ml are serial diluted at 1:3 in PBS
containing 1% BSA using 96-well plate100 pl of the appropriate chimeric TrkA-B-transfected 293 cells or control parental 293 cells at 2x10e6 cells/ml in serum-free growth medium are added to U-bottom 96 well plate to get 1x1 0e5 cells/well.
The cells are centrifuged down at 1600cpm for 2minutes. The supernatants are discarded with one-time swing and the plate is patted gently to loosen the cell pellet.
100 pl of diluted primary anti-TrkB or isotype-matching control antibodies in cold PBS
containing 10 % FCS are added to the cells and incubated on ice for 1 hour.
The cells are then stained with 100 pl of diluted secondary anti-IgG antibody HRP
conjugates (Donkey anti-Rabbit IgG, Thermo,cat # 31458; goat anti-mouse IgG FC, Pierce, 31439; Goat anti-human IgG Fc, Pierce, cat # 31413) on ice for 1 hour.
Following each step of primary antibody and secondary antibody incubations, the cells are washed 3 times with ice-cold PBS100 pl of substrate TMB1 component (B10 FX, TMBW ¨0100-01) is added to the plate and incubated for 10 minutes at room temperature. The color development is stopped by adding 100 pl of 0.18M H2SO4.

- 177 -The cells are centrifuged down and the supernatants are transferred to fresh plate and read at 450 nm (Soft MAX pro 4.0, Molecular Device).
As shown in Figure 33, in this cell-based ELISA the anti-TrkB antibody TOA-1 binds to cell-surface expressed TrkB (sequence 34), but not TrkA (sequence 33). In this same assay format, TOA-1 binds to chimeric TrkA (d5TrkB), but not TrkB
(d5TrkA). Additionally, TOA-1 binds to TrkB (d4TrkA) and not to TrkA (d4TrkB).

Together the results indicate that the TOA-1 antibody binds to domain 5 of TrkB, and domain 4 of TrkB is not sufficient for TOA-1 binding.
Example 51 Isolation of cat NTRK2 (TrkB) cDNA
Methods Isolation of Cat (Fells domesticus ) TrkB
TrkB Cat coding sequences were isolated and cloned using standard Polymerase Chain reaction (PCR) methods. Full length cat (Fells domesticus) TrkB
sequences were amplified from a cat brain cDNA pool (BioChain) using Stratagene Easy-A High-Fidelity system (cat# 600640) and the suggested protocol using the oligonucleotides, 5'GGATCCGCCGCCACCATGTCGTCCTGGACGAGGTGGCATGG
(SEQ ID NO:144) and 5'GCGGCCGCCTAGCCCAGAATATCCAGGTAGACCGGAGAT (SEQ ID NO:145), as primers. The cDNA was cloned into pCR2.1-TOPO vector (Invitrogen) and subsequently subcloned into pcDNA3.1-Hyg (Invitrogen) with BamHI and Notl restriction enzyme sites. The resultant plasmids were sequenced (SEQ ID
NO:140, 141) Dog (Canis familiaris) full length TrkB (XM_851329) coding sequence was amplified by PCR from a dog brain cDNA pool (BioChain) as described above using the oligonucleotides 5'GGATCCGCCGCCACCATGTCGTCCTGGACGAGGTGGCATGG (SEQ ID
NO:146) and 5'GCGGCCGCCTAGCCTAGAATATCCAGGTAGACTGGAG (SEQ ID
NO:147), as primers. The dog ortholog of human TrkB isoform c was selected for subcloning into pcDNA3.1-Hyg as described above. The resultant plasmids were sequenced (SEQ ID NO:142, 143) = CA 02891714 2015-05-15

- 178 -Example 52 Antibody binding to TrkB of different species was measured by cell-based ELISA
Cell-based ELISA was performed to evaluate the binding of anti-TrkB antibodies to mouse (nm 001025074), cat, and dog TrkB receptors Cell transfection:
Human embryonic kidney 293 cells (ATCC) cells are plated at 5x10^6 cells per 10cm2 tissue culture plate and cultured overnight at 37C The next day cells are transfected with Human, dog, or cat TrkB expression plasmids using Fugene6 (Roche Applied Sciences) at a 3:1 ratio of reagent to plasmid DNA using the manufacturer's protocol. Cells are harvested 24hrs after transfection using Accutase (Millipore), washed once with phosphate buffered saline (PBS), then suspended in DMEM with 0.2% BSA at 2x10e6 cells/ml.
Cell-based ELISA assay Anti-TrkB or control antibodies at 10 pg/ml are serial diluted at 1:3.17 in DMEM containing 0.2% BSA using 96-well plate50 pl of the appropriate TrkB-transfected 293 cells or control LacZ-transfected 293 cells from the above are added to a U-bottom 96 well plate to get 1x1 0e5 cells/well. The plate is left at 4C
for 15 min before 50 pl of the diluted primary anti-TrkB or isotype-matching control antibodies are added to the cells. The cells and antibody are mixed by gentle pipetting then incubated at 4C for 1 hour. The cells are washed 3 times with ice-cold PBS by centrifugation at 1600cpm for 2 minutes. Each time the supernatants are discarded with one-time swing and the plate is patted gently to loose the cell pellet before adding the next buffer or medium Then, 100 pl of diluted secondary anti-IgG
antibody HRP conjugates (Pierce) in DMEM with 0.2% BSA is added to the cells. Cells are incubated at 4C for 1 hour, and washed 3 times as above. For staining, 100 pl of substrate TMB1 component ( BIO FX, TMBW ¨0100-01) is added to each well and incubated for 5-30 minutes at room temperature. The color development is stopped by adding 100 pl of 0.18M H2SO4 The cells are centrifuged down and the supernatants are transferred to fresh plate and read at 450 nm (Soft MAX pro 4.0, Molecular Device).

- 179 -The anti-TrkB antibodies mouse TOA-1 and humanized TOA-1 bind to mouse, cat and dog TrkB as determined by cell-based ELISA, shown in Figure 33. The values for the binding of TOA-1 to cell surface mouse, dog, and cat TrkB, as determined by cell-based ELISA, are shown in Table 31.
Table 31 EC50 values for binding to TrkB determined by cell-based ELISA
Humanized TOA1 Mouse TOA1 293-mouse TrkB 0.33 nM 0.40 nM
293-cat TrkB 0.52 nM 0.23 nM
293-dog TrkB 0.55 nM 0.65 nM

- 180 -Example 53 Selectivity of Anti-TrkB antibodies against TrkA, TrkC, and p75 Multiple experimental approaches were used to demonstrate that the anti-TrkB
antibody, TOA-1, is selective for human TrkB versus human TrkA, TrkC, and the low affinity BDNF receptor p75 NTR.
Direct binding ELISA
TOA-1 selectivity to TrkB was assessed by direct binding ELISA to recombinant human TrkA-Fc, TrkB-Fc or TrkC-Fc with biotinylated humanized TOA-1, chimeric TOA-1 and mouse TOA-1 antibodies as follows.
96-well plates (Costar) were coated with 1 pg/ml rhTrkB-ECD (R&D system, 688-TK), 5 pg/ml rhTrkA-ECD (R&D system, 175-TK), or 5 pg/ml rhTrkC-ECD (R&D
system, 373-TUTF) in PBS and incubated overnight at 4 C Plates were blocked with PBS + 0.2% casein (100 pl per well) for 3 hours at room temperature Next, 100 pl of biotinylated antibody (murine TOA1, chimeric TOA1, humanized TOA-1, or isotype control) at 6.7 nM was added to the wells and incubated for 1 hr at room temperature The wells were washed four times with PBS + 0.03% tween-20 Streptavidin-HRP
(Southern Biotech catalog #7100-05) diluted 1:10,000 was added and incubated for 30 minutes at room temperature The wells were washed four times with PBS +
0.03% tween-20 and TMB (BioFx) was added The reaction developed for 5-10 minutes and was then quenched with 0.18 N H2SO4The absorbance at 450nm was determined As shown in Figure 34, humanized TOA-1 (i.e., TAM-163), chimeric TOA-1 and mouse TOA-1bind TrkB-Fc but not TrkA-Fc or TrkC-Fc.
FACs analysis 90% confluent HEK293 cells were transiently transfected with plasmids expressing human TrkB (open reading frame from nm_006180 cloned into the mammalian expression vector pcDNA3.1-hyg, Invitrogen) or human p75NTR (open reading frame from NM_002507 cloned into vector pSMED2) using Fugene6 (Roche Applied Sciences) according to the manufacturer's directions Expression of human TrkB and human p75NTR was verified by Western analysis. At 24 hours post

- 181 -transfection, the cells were harvested, washed with PBS, resuspended in PBS/
0.5 %
BSA 2.5x10^5 huTrkB and hu p75NTR cells were stained with antibodies as followsFor p75NTR detection, cells were incubated with 1 ug/ml mouse anti-P75-A1exa488 (Millipore MAB5368X) for 30min at 4 C, followed by a washing with PBS
through centrifugation (1500 rpm for 5 min). For TrkB staining, cells were incubated with 1 pg/ml humanized TOA-1 antibody for 30min at 4 C followed by a PBS wash as described aboveNext, the cells were incubated with FITC labeled mouse anti-Human IgG (Southern Biotech S9670-02) for 30min at 4 C, followed by a washing with PBS
as described above Stained cells were analyzed on a FACSCalibur using CellQuest software (Becton Dickinson) As shown in Figure 35, using FACS analysis TOA-1 antibody binds to cell-surface expressed human TrkB, but does not bind to cell-surface expressed human p75NTR FACS analysis using anti-p75 on the respective cell lines confirms the cell-surface expression of p75 in these cell lines.
Cell-based ELISA
Cell-based ELISAs were performed to evaluate the binding of TOA-1 to human TrkB receptor but not to p75NTR.
Human embryonic kidney 293 cells were transfected and harvested as above except resuspended in DMEM containing 0.2% BSA at 2x10e6 cells/ml Anti-TrkB
(TOA-1), control antibodies (anti-p75NTR, R&D AF367) or anti-hulgG isotype antibody, at 20 pg/ml are serial diluted at 1:3.17 in DMEM containing 0.2% BSA
using 96-well plate50 pl of the transfected or control 293 cells from the above are added to a U-bottom 96 well plate to get 1x1 0e5 cells/well. The plate is incubated at 4C for 15 min before 50 pl of the diluted primary anti-TrkB or isotype-matching control antibodies in cold are added to the cells. The cells and antibody are mixed by pipetting up-down three times before incubated at 4C for 1 hour. The cells are washed 3 times with ice-cold PBS by centrifugation at 1600cpm for 2 minutes.
Each time the supernatants are discarded with one-time swing and the plate is patted gently to loose the cell pellet before adding the next buffer or medium Then, 100 pl of diluted secondary anti-IgG antibody HRP conjugates (Pierce) in DMEM with 0.2%

- 182 -BSA is added to the cells. Cells are incubated at 4C for 1 hour, and washed 3 times as above. For staining, 100 pl of substrate TMB1 component ( BIO FX, TMBW ¨
0100-01) is added to each well and incubated for 5-30 minutes at room temperature.
The color development is stopped by adding 100 pl of 0.18M H2SO4The cells are centrifuged down and the supernatants are transferred to fresh plate and read at 450 nm (Soft MAX pro 4.0, Molecular Device).
As shown in Figure 36, the TOA-1 antibody binds to the cells expressing TrkB
but not to the cells transfected with p75NTR construct (Fig 36A), which does express p75NTR on the cell surface as detected by anti-p75NTR antibody (Fig 36B) Trk signal transduction The ability of mouse, chimeric, and humanized anti-TrkB TOA-1 antibodies to activate the TrkB signaling cascade, but not the TrkA or TrkC cascades, was assessed by monitoring autophosphorylation of Trk and phosphorylation of ERK1/2, AKT, and PLCy1, known mediators of Trk signaling (reviewed in Friedman et al Exp Cell Res 1999; 253:131-142) Stable cell lines of HEK-293 cells expressing both a CRE-luciferase reporter and rhuTrkA (open reading frame from NM_002529.3), rhuTrkB (open reading frame from nm_006180), and rhuTrkC (open reading frame from NM_001012338.1) were generated using standard techniques (Zhang et al, 2007, Neurosignals 15:29-39) Stable cell lines are designated rhuTrIcA-CRE, rhuTrkB-CRE, and rhuTrkC-CRE.
Trk-expressing cells were treated with TOA-1 antibodies, isotype control antibodies, or the neurotrophin ligands BDNF, NGF, and NT3 Western analyses was performed to evaluate autophosphorylation of hTrk (Tyr490) and phosphorylation of ERK1/2 (Thr202/Tyr204), AKT (Ser473), p38 (Thr180/Tyr182) and PLCy1 (Tyr783) as detailed below.
rhuTrkA-CRE, rhuTrkB-CRE, rhuTrkC, or parental HEK-293 cells were plated in 6-well plates at 5x 105 cells/well in 10% FCS ¨ DMEM growth media and cultured until the cells were 85-90% confluent. Cells were washed once with 0.1% FCS ¨
DMEM (low serum media) and incubated for an additional 4 hours in low serum media. Next, cells were treated with TOA-1 antibody (final concentration 100nM) or

- 183 -neurotrophin (TrkB: BDNF,R&D #248BD 10nM final concentration; TrkA: NGF, R&D
256GF 10nM final concentration; TrkC: NT3, R&D267N3, 25nM final concentration) for 15-60 min. Medium was aspirated from the wells and 0.6 ml of 1X loading buffer (Invitrogen NP0007, with 1% b-ME) was added per well to lyse the cells. Cell lysates were transferred to Eppendoff tubes, and heated at 100 C for 5 min 25 pl of each sample were resolved on a NuPAGE 4-20% Bis-Tris gradient gel (Invitrogen) Western analysis was performed as follows. After electrophoresis, size-fractionated proteins were transferred onto nitrocellulose membranes Membranes were blocked with 5% milk in T-TBS (0.15% Tween20 in TBS), incubated with the appropriate primary antibody [anti-P-Irk: Phosphor-TrkA (Tyr490),Cell signaling (CS) #9141; anti-Phospho-PLCy1 (Tyr783), CS #2821; anti-Phospho AKT (Ser473), CS#9271; anti-P-ERK1/2 (Phospho-P44/P42 (Thr202/Tyr204), CS #9101; anti-Actin, Sigma A2066] in 1% milk T-TBS on a rocking platform at 4 C overnight Membranes were washed 3X in T-TBS, then incubated with the appropriate HRP-conjugated secondary antibody (Cell Signaling #7974) for 2 hours. Next, membranes were washed 4 times in T-TBS and once in TBS. The signals were developed using ECL
kit (GE RPN2106V) and the manufacturer's protocol followed by x-ray film exposure or Gel-Doc (Bio-Rad) to capture the image.
As shown in Figure 37, TOA-1 antibodies failed to cause a detectable increase above basal levels in Trk autophosphorylation, and PLCy1, AKT, and ERK1/2 phosphorylation in TrkA and TrkC expressing cells. In contrast, the TrkA and TrkC
ligands, NGF and NT3, respectively, induced a response indicating that the cellular signaling system is intact.
In all further examples, huT0A-1 (SEQ ID NOs: 51 and 53), is referred to as TAM-163.
Example 54 mRNA expression of catalytic and non-catalytic TrkB isoforms In preparation for functional studies and to identify tissues and cell lines expressing high levels of endogenous TrkB, the tissue distribution of catalytic compared to non-catalytic isoforms of TrkB was examined using Taqman quantitative PCR (Q-PCR). Primer-probe pairs were designed to recognize either the extracellular , CA 02891714 2015-05-15

- 184 -domain (ECD) common to all hTrkB isoforms or the catalytic domain common to the catalytic hTrkB-a and hTrkB-c isoforms. A standard curve was generated for each primer probe pair using TrkB plasmid cDNA, and was used to convert raw data into TrkB cDNA molecules. Assuming similar efficiency of reverse transcription for different mRNA samples, this number reflects the molecules of TrkB mRNA for each tissue. Two independent primer-probe pairs were designed for each region and similar results were obtained with both pairs. As can be seen in Table 32, hTrkB is most highly expressed in the brain, and in this tissue the catalytic isoforms of TrkB
accounts for - 35% of all TrkB isoforms. The neuroblastoma cell line SH-SY5Y, when differentiated with retinoic acid, expresses levels of TrkB mRNA
comparable to the ones found in human brain with 87% of the TrkB mRNA accounted for by the catalytic isoforms. This cell line was therefore chosen to evaluate the effects of TAM-163 on endogenous TrkB. Non-neuronal tissues showed <10% of the TrkB mRNA levels found in brain, when examining expression of all isoforms; expression of the catalytic isoforms was even lower and constituted <2% of the amount observed in brain.
The lowest expression of TrkB was observed in peripheral blood leukocytes where TrkB
mRNA was barely detectable.
Table 32 mRNA expression of TrkB in normal human tissues - comparison of catalytic isoforms versus total TrkB
Catalytic Isoforms All Isoforms (total) Ratio Human Tissue Molecules/PC
% of brain Molecules/PC % of brain catalytic/tot R R al (%) Brain 457311 100.00 1363034 100.00 33.55 Kidney 8849 1.94 116365 8.54 7.60 Artery 4780 1.05 63754 4.68 7.50 Skin 4079 0.89 113506 8.33 3.59 Uterus 3130 0.68 59566 4.37 5.26 Heart 2710 0.59 74769 5.49 3.62 Vein 2666 0.58 101286 7.43 2.63 Adipose Tissue 2183 0.48 67345 4.94 3.24 Ovary 1948 0.43 59626 4.37 3.27 Pancreas 1827 0.40 58539 4.29 3.12 Thymus 1246 0.27 33902 2.49 3.68 Spleen 1071 0.23 50565 3.71 2.12 Skeletal Muscle 777 0.17 12341 0.91 6.29

- 185 -Table 32 mRNA expression of TrkB in normal human tissues - comparison of catalytic isoforms versus total TrkB
Catalytic lsoforms All lsoforms (total) Ratio Human Tissue Molecules/PC %
of brain Molecules/PC % of brain catalytic/tot al (%) Lymph Node 774 0.17 24328 1.78 3.18 Colon 677 0.15 16491 1.21 4.10 Bone Marrow 411 0.09 13101 0.96 3.14 Lung 404 0.09 11215 0.82 3.60 Ileum 308 0.07 7048 0.52 4.38 Jejunum 173 , 0.04 4148 0.30 4.18 Stomach 114 0.03 2626 0.19 4.36 Liver 101 0.02 2672 0.20 3.76 White Blood 81 0.02 40 0.00 N/A
Cells (Buffy _ Coat) Duodenum 63 0.01 1878 0.14 3.35 Peripheral 5 0.00 30 0.00 N/A
Blood Leukocytes SH-SY5Y Cells 908417 198.64 1042645 76.49 87.13 (differentiated) SH-SY5Y Cells 3318 0.73 6056 0.44 54.78 (undifferentiate d) Taqman quantitative PCR conditions, primer-probe pairs and conversion into molecules of cDNA/200 ng RNA is described in Error! Reference source not found. ; the average value for the two primer-probe pairs available for catalytic and total trkB isofornns is shown. A ratio of catalytic/total was not calculated if the values for total were too low for accurate quantitation (indicated as N/A).
Example 55 Signaling by TAM-163 in cell lines expressing recombinant hTrkB, hTrkA and hTrkC
The ability of TAM-163 to activate the TrkB signaling cascade was assessed using 1) a transcriptional reporter assay to monitor TrkB signaling activation, 2) an enzyme complementation assay to monitor recruitment of the signaling molecule SHC1 to TrkB and 3) evaluating autophosphorylation of hTrkB and phosphorylation of ERK1/2, AKT, and PLCyl, known mediators of TrkB signaling. The same assays were also used to examine the ability of TAM-163 to activate TrkA and TrkC
signaling pathways.

- 186 -Example 56 TAM-163 activates the Cre-luciferase reporter gene in cell lines expressing hTrkB, but not in cell lines expressing hTrkA or hTrkC
The Cre-luciferase (Cre-luc) transcriptional reporter assay measures the ability of TrkB ligands to activate the CRE response element and as such integrates multiple upstream signaling pathways. The cell lines used for this assay, hTrkB-Cre, hTrkA-Cre and hTrkC-Cre, have been previously described and were shown to respond specifically to the appropriate endogenous ligands (Zhang et al. Neurosignals.

2007;15(1):26-39, Qian et al. J Neurosci. 2006 Sep 13;26(37):9394-9403).
Treatment of hTrkB-Cre cells with TAM-163 resulted in a dose-dependent increase in luciferase activity; the EC50 was 0.2 nM and the maximum fold-increase was 5-fold (Figure 38).
In the same experiment, human BDNF (Peprotech, Rocky Hill, NJ) showed an EC50 of 5.2 nM and a maximum fold-increase of 7.5-fold, while a hIgG control antibody had no effect (Figure 38). Between experiments, EC50 and maximum-fold induction ranged from 0.10 nM - 0.79 nM and 2.5-5.4 fold for TAM-163, and from 5.2 - 8.2 nM
and 6-7.5-fold for human BDNF. Individual activity data for different lots of tested in the hTrkB Cre-luc reporter assays multiple times are shown below (Table 33). Averaging across all assays, the EC50 value for TAM-163 in this assay was determined to be 0.37 0.06 nM with a fold-induction of 4.2 0.3.
Table 33. Activity of different lots of TAM-163 in the hTrkB Cre-luc transcriptional reporter assay TAM-163 Lot# EC50 (nM) Fold-increase Reference L40042-166 0.34 2.5 L401310-192 L40042-166 0.65 3.0 L401310-207 L40042-166 0.35 3.1 L42358-41 L40042-166 0.45 5.4 L42358-41 L40042-166 0.79 5.0 L42358-52 L40042-166 0.2 4.5 L42358-173 L40042-192 0.24 2.9 L42358-41 L40042-192 0.39 5.3 L42358-41 L40042-192 0.61 5.0 L42358-52 L42385-008 0.25 2.4 L42358-41 L42385-008 0.37 5.3 L42358-41 L42385-024 0.13 5.3 L42358-195 L42385-152 0.10 4.5 L42358-173 Average SEM 0.37 0.06 4.2 0.3 , CA 02891714 2015-05-15

- 187 -The crossreactivity of TAM-163 with human TrkA and TrkC was tested using hTrkB-Cre, hTrkA-Cre and hTrkC-Cre with the appropriate endogenous controls (NGF for hTrkA, BDNF for hTrkB and NT-3 for hTrkC) and 20 nM or 100 nM TAM-163 (Figure 39). While the endogenous ligands resulted in the expected responses (NGF: 4.5-fold increase hTrkA-Cre; NT-3: 2.9-fold increase hTrkC-Cre), TAM-163 did not show any increased luciferase activity compared to hIgG control in either hTrkA
or hTrkC cells. In the same experiment, TAM-163 did activate hTrkB-Cre cells 3.4 -4.1-fold, demonstrating that TAM-163 was effective on TrkB, as expected.
Example 58 TAM-163 mediates SHC1 recruitment in cells expressing hTrkB, but not in cells expressing hTrkA or hTrkC
The ability of TAM-163 to activate hTrkA, hTrkB and hTrkC was also assayed in an enzyme complementation assay (Figure 40;Table 34). This assay monitors recruitment of the signaling molecule SHC1 to autophosphorylated TrkB in U2OS
cells using the Discoverx Pathhunter technology. TAM-163 activated hTrkB in this assay with a potency similar to the one observed in the Cre-luciferase reporter assay (EC50= 0.67 nM) (Figure 40;Table 34). Similar to the Cre-luciferase assay, the maximum signal induced by TAM-163 was significantly less than the maximum signal observed with BDNF, suggesting that TAM-163 is a partial agonist in this assay.
Importantly, TAM-163 did not activate hTrkA or hTrkC at concentrations up to nM, while the endogenous ligands for these receptors showed strong activation at very low concentrations (Figure 40;Table 34). hTrkA, hTrkB and hTrkC were activated by an unusually broad array of endogenous Trk ligands in this assay which could reflect the particular signaling pathway assayed (SHC1 recruitment), the cell background (U2OS cells) or a peculiarity of the Pathhunter system (this assay uses the hTrkB-a isoform fused to a small peptide epitope, while our other assays are conducted with the native hTrkB-c isoform). It is remarkable that TAM-163 did not crossactivate hTrkA and hTrkC despite the apparent relaxed specificity for the endogenous ligands in this assay.

- 188 -Table 34. Activity of TAM-163 on hTrkA, hTrkB and hTrkC in the SHC1 recruitment assay TAM-163 hIgG NGF BDNF NT-3 NT-4 hTrkA EC50 (nM) N/A N/A 0.48 30.13 1.49 0.43 Fold-increase 1.1 1.1 7.2 4.3 9.2 9.0 hTrkB EC50 (nM) 0.67 N/A 1.59 1.02 0.50 0.33 Fold-increase 5.2 1.1 7.2 5.9 7.7 6.1 hTrkC EC50 (nM) N/A N/A N/A 5.45 0.86 N/A
Fold-increase 1.0 1.0 1.9 8.4 10.5 1.1 The maximum ligand concentration tested was 670 nM (antibodies) and 74 nM
(endogenous ligands).
The EC50 and maximum fold-increase reached is shown. EC50 = concentration at which 50% of maximum effect is reached. An EC50 was not calculated if a plateau was not reached (indicated as N/A).
Example 59 TAM-163 activates Trk-dependent phosphorylation events in cells expressing hTrkB, but not in cells expressing hTrkA or hTrkC
To directly monitor signaling events downstream of TrkB, we used Western blotting. Autophosphorylation of TrkB (Y490) as well phosphorylation of signaling molecules downstream of TrkB, including ERK1/2 (Thr202/Tyr204), PLCyl (Tyr783) and AKT (Ser473) were assessed using the hTrkA-Cre, hTrkB-Cre and hTrkC-Cre cell lines described above. TAM-163, but not a hIgG control antibody, induced dose-dependent phosphorylation of TrkB (Y490), ERK1/2 (Thr202/Tyr204), PLC71 (Tyr783) and AKT (Ser473) in hTrkB-Cre cells (Figure 42).
TAM-163 was unable to induce signaling in hTrkA-Cre or hTrkC-Cre cells (Figure 42). The endogenous ligands for hTrkA (NGF) and hTrkC (NT-3) induced both Trk autophosphorylation and phosphorylation of the signaling intermediates ERK1/2 and PLC71 in hTrkA-Cre and hTrkC-Cre cells, demonstrating that these cell lines respond to their appropriate ligand. In the same experiment, TAM-163 was able to activate signaling downstream of hTrkB, demonstrating that TAM-163 was active.
Example 60 TAM-163 activates Trk-dependent phosphorylation events in a human neuroblastoma cell line expressing endogenous hTrkB
To examine the ability of TAM-163 to signal in cells expressing endogenous TrkB, we used differentiated human neuroblastoma SH-SY5Y cells. TAM-163, but not

- 189 -a hIgG control antibody, induced phosphorylation of ERK1/2, PLCy1, and AKT in a dose-dependent manner in these cells; effects became apparent at concentrations 1 nM TAM-163 (Figure 43). Compared to BDNF, TAM-163 appeared somewhat less potent and showed a significantly lower maximal stimulation of phosphorylation, suggesting that TAM-163 is a partial agonist in this system.
Example 61 TAM-163 induces internalization and degradation of hTrkB
BDNF has been reported to mediate internalization and degradation of TrkB
{Error! Reference source not found.,Error! Reference source not found.}. We examined the effect of TAM-163 on TrkB internalization and degradation using cell lines expressing either recombinant (hTrkB-Cre) or endogenous (SH-SY5Y) TrkB.
To monitor internalization, cells were activated with TAM-163 or BDNF for the indicated times, cell surface proteins were then labeled with biotin, isolated by strepatavidin affinity purification and cell surface TrkB protein was identified by Western blotting. In this assay, biotinylated TrkB represents the TrkB remaining on the cell surface after activation. As can be seen in Figure 44, TAM-163, but not a control hIgG
antibody, induced significant internalization of TrkB in cells expressing recombinant TrkB
(hTrkB-Cre) and in cells expressing endogenous TrkB (SH-SY5Y). TAM-163 did not affect cell surface levels of unrelated proteins (EGF-receptor, NMDA-receptor). The time-course and amount of TrkB internalization induced by TAM-163 was comparable to BDNF in both hTrkB-Cre and SH-SY5Y cells (Figure 44).
To monitor TrkB degradation, cell surface proteins were labeled with biotin prior to ligand exposure, and cells were then activated with TAM-163 or BDNF
for the indicated times. Labeled proteins were isolated by strepatavidin affinity purification, and TrkB was identified by Western blotting. In this assay, biotinylated TrkB
represents the total TrkB remaining after activation; the disappearance of labeled TrkB is a measure for its clearance from the cell. As can be seen in Figure 45, TAM-163, but not a hIgG control antibody, induced degradation of cell-surface-labeled TrkB in cells expressing recombinant TrkB (hTrkB-Cre) and in cells expressing endogenous TrkB (SH-SY5Y). TAM-163 did not affect unrelated proteins (EGF-receptor, NMDA-receptor). The time-course and amount of TrkB degradation induced . CA 02891714 2015-05-15

- 190 -by TAM-163 was comparable to BDNF in both hTrkB-Cre and SH-SY5Y cells (Figure 45).
Example 62 TAM-163 does not bind to human p75NTR
The crossreactivity of TAM-163 to human p75NTR was examined by fluorescence activated cell sorter (FACS) analysis using HEK293 cells transiently transfected with hTrkB or human p75NTR. TAM-163 (6.7 nM) was able to specifically bind to cells transfected with hTrkB, as evidenced by increased fluorescence compared with cells transfected with control empty vector (Figure 46). TAM-163 did not show any binding to cells transfected with human p75NTR; in fact, staining was slightly less compared to control transfected cells (Figure 46, top panel). To verify that p75NTR was indeed expressed and present at the cell surface, cells were stained with an ALEXA-labelled anti-p75NTR antibody. As can be seen in Figure 46, bottom panel, the anti-p75NTR antibody strongly stained cells expressing p75NTR, but did not stain cells expressing hTrkB or a control vector.
As a second approach, we tested the binding of TAM-163 to cells expressing p75NTR using a cell-based ELISA. HEK293 cells transiently transfected with hTrkB, human p75NTR or control vector were incubated with either TAM-163 or anti-p75NTR antibody. TAM-163 specifically bound to cells expressing human TrkB
with binding detectable at concentrations as low as 0.2 nM (Figure 47). TAM-163 did not show any binding to cells expressing human p75NTR even at very high concentrations (67 nM). To verify that p75NTR was indeed expressed and present at the cell surface, staining with anti-p75NTR antibody was used. As expected, anti-p75NTR antibody stained cells expressing human p75NTR, but did not stain control cells or cells expressing hTrkB (Figure 47).
Example 63 Crossreactivity of TAM-163 with monkey, mouse, dog and cat TrkB
Since no sequence information is available for cynomolgus monkey TrkB, we isolated TrkB cDNA from this species using standard cloning biology techniques and brain as a template. Sequencing revealed the presence of both TrkB-c and TrkB-a isoforms with the majority of clones (8/10) containing the TrkB-c isoform.
Comparison , CA 02891714 2015-05-15

- 191 -of cynomolgus TrkB cDNA sequence with the human TrkB sequence shows that, with the exception of one amino acid change in the signal sequence, the mature cynomolgus monkey TrkB protein is identical in amino acid sequence to human TrkB.
The rhesus monkey TrkB sequence (available in public databases as XP 001107264) is found to be identical to mature human TrkB (not shown). Since the monkey TrkB protein is identical to human, all the human TrkB binding and signaling data shown above are equally applicable to monkey TrkB.
We evaluated the crossreactivity of TAM-163 with mouse and dog TrkB using both binding and signaling experiments. For binding, a cell-based ELISA was used.
HEK293 cells transiently transfected with mouse, dog or cat TrkB, or a control vector, were incubated with various concentrations of the TAM-163 antibody. Dose-dependent binding was observed to TrkB from all species, while no binding was observed to a control cell line expressing lacz. The EC5Os were similar between species (mouse TrkB = 0.34 nM; dog TrkB = 0.94 nM; cat TrkB = 0.39 nM), indicating that TAM-163 binds to mouse, dog and cat TrkB with high affinity (Figure 48).
To evaluate the ability of TAM-163 to induce signaling on mouse and dog TrkB, we generated stable cell lines expressing either mouse or dog TrkB. TAM-163, but not a hIgG control antibody, dose-dependently activated phosphorylation of TrkB
(Y490), ERK1/2 (Thr202/Tyr204), PLC71 (Tyr783) and AKT (Ser473) in both mouse (Figure 49; TAM-163) and dog (Figure 50; TAM-163) TrkB cells. Activation of signaling pathways was detectable at concentrations __ 1 nM in both mouse and dog, consistent with the observed EC50 in the binding assay for mouse and dog TrkB.
Example 64 TrkB in sensory neuronal hearing loss:
Temporary and permanent hearing loss is induced by various sources including overexposure to intense sound, chemo induced damage or neurodegeneration that occurs from aging (presbycusis). Recent evidence from Liberman, 2009, J. Neurosci. 29(45):14077-14085, suggests that the ribbon synapse is the first site of insult after both acute noise exposure and in presbycusis. This ribbon synapse damage preceeds spiral ganglia neuron (SGN) and hair cell loss, rendering the ribbon synapse an attractive target for hearing loss intervention.

. CA 02891714 2015-05-15

- 192 -It is known that factor-related peptides known as neurotrophins are essential for neural development and maintenance, and that several neurotrophins act on a number of neuronal receptors, promoting neuronal survival and differentiation.
This class of peptides has been shown to impact the ribbon synapse, specifically the brain-derived neurotrophic factor (BDNF). The elegant work of Wise et al (J.
Comp.
Neurol. 2000, 487:147-165, whose contents are hereby incorporated by reference), clearly demonstrates that BDNF treatment prevented the loss of SGNs in response to deafening consistently across all cochlear regions. Specifically, application of BDNF
prevented auditory neuron death, reduced continued neuronal loss, and enhanced cochlear performance in the models tested, producing a profound effect on hearing.
Meister et al (Curr Biol, 2014: 24(6): 658-663, whose contents are hereby incorporated by reference) demonstrated TrkB mediated protection against circadian sensitivity to noise trauma in murine cochlea. Schimmang et al (Development, 2003, 130: 4741-4750, whose contents are hereby incorporated by reference) demonstrated that a lack of BDNF and TrkB signaling in the postnatal cochlea leads to a spatial reshaping of innervation along the tonotopic axis and hearing loss.
Tyrosine Kinase Receptor B (TrkB) is a high affinity catalytic receptor for several growth factor-related peptides (neurotrophins), in particular BDNF and neurotrophin-3 (NT-3). TrkB is expressed and functions predominantly in neurons throughout the central nervous system, including the ribbon synapse rendering it a strong potential therapeutic target for a variety of sensorineural hearing loss disorders including sudden hearing loss, noise induced hearing loss, age related hearing loss (presbycusis), noise induced hearing loss, drug induced hearing loss and genetic disorders of hearing. Thus, TrkB agonists may be potential therapeutics for treatment of such hearing loss disorders.
TAM-163 is a humanized monoclonal antibody designed as an agonist of the TrkB. TAM-163 (also referred to as huT0A-1 and PF-05230901) has been shown to be such a selective antibody TrkB agonist showing strong activation of downstream signaling cascades (Figure 41, TRK, PLCy1 and ERK 1/2) and demonstrating an excellent selectivity profile (Figure 42). Thus, TAM-163 may be a potential

- 193 -therapeutic for treatment of hearing loss disorders that may be treated by a TrkB
agonist in a patient in need of such treatment.
Such patients can be identified by a test to determine hearing loss which may be conducted by an audiologist using an audiometer to determine the individual's hearing sensitivity at different frequencies. Other hearing tests may be used, for example, the Weber test, the Rinne test, the Hearing in Noise test, the acoustic reflex test, and a tympanogram, among many such tests known in the art.
Example 65 Conclusions The present data demonstrate that TAM-163 is a potent and specific agonist of human TrkB that activates all aspects of the TrkB signaling cascade. While the potency (EC50) of TAM-163 is comparable to the endogenous TrkB ligand BDNF, the maximum effect is less than what is observed with BDNF (-50-80% of maximal signal depending on the assay), suggesting that TAM-163 is a partial agonist of human TrkB. TAM-163 induces internalization and degradation of human TrkB in a manner similar to BDNF. TAM-163 does not crossreact with human TrkA, human TrkC or human p75NTR in cell-surface binding experiments and does not induce signaling in cell lines expressing human TrkA or TrkC. TAM-163 binds to and activates mouse and dog TrkB at low nanomolar concentrations, similar to its effect on human TrkB.
Since monkey TrkB is 100% identical to human TrkB, TAM-163 also fully crossreacts with monkey TrkB. Examination of the mRNA expression of the catalytic isoforms of TrkB as well as all TrkB isoforms in normal human tissues confirms that the catalytic isoform of TrkB is most highly expressed in the brain and that the human neuroblastoma cell line SH-SY5Y can be used to examine signaling mediated by endogenous TrkB.
FURTHER EMBODIMENTS OF THE INVENTION
E77. According to the 77th embodiment of the invention (E77), there is provided an isolated Human Tyrosine Receptor Kinase B (huTrkB) antibody which specifically binds to huTrkB wherein the VH region comprises an amino acid sequence selected from the group consisting of SEQ ID NO: 51, SEQ ID NO:111, and SEQ ID NO:113, . CA 02891714 2015-05-15

- 194 -and wherein the VL domain comprises an amino acid sequence selected from the group consisting of SEQ ID NO:53, SEQ ID NO:115, SEQ ID NO:117, SEQ ID
NO:119, SEQ ID NO:121, SEQ ID NO:124, SEQ ID NO:126, SEQ ID NO:128, SEQ
ID NO:130, and SEQ ID NO:132.
E78. The antibody as set forth in E77, wherein the VH region comprises the amino acid of SEQ ID NO:51 and the VH region comprises the amino acid sequence of SEQ
ID NO:53.
E79. The antibody as set forth in any one of E77-E78, wherein the antibody is IgG1 subclass.
E80. The antibody as set forth in any one of E77-E79, wherein the HC comprises SEQ ID NO:75 and the LC comprises SEQ ID NO:78.
E81. An isolated Human Tyrosine Receptor Kinase B (huTrkB) antibody which specifically binds to huTrkB wherein the VH region comprises an amino acid sequence encoded by sequence selected from the group consisting of SEQ ID
NO:110, SEQ ID NO:111, and SEQ ID NO:113; and wherein the wherein the VL
region is encoded by a nucleic acid comprising a sequence selected from the group consisting of SEQ ID NO:123, SEQ ID NO:116, SEQ ID NO:118, SEQ ID NO:120, SEQ ID NO:122, SEQ ID NO:125, SEQ ID NO:127, SEQ ID NO:129, SEQ ID
NO:131, and SEQ ID NO:133.
E82. An antibody as set forth in any one of E77-E81, wherein the VH region is encoded by a nucleic acid comprising SEQ ID NO:110, and the VL region is encoded by a nucleic acid comprising SEQ ID NO:123.
E83. A nucleic acid encoding the antibody as set forth in any one of E77-E82.
E84. A nucleic acid encoding the antibody as set forth in any one of E77-E82, wherein the VH region is encoded by a nucleic acid comprising a sequence selected from the group consisting of SEQ ID NO:110, SEQ ID NO:111, and SEQ ID NO:113;
and wherein the wherein the VL region is encoded by a nucleic acid comprising a sequence selected from the group consisting of SEQ ID NO:123, SEQ ID NO:116, SEQ ID NO:118, SEQ ID NO:120, SEQ ID NO:122, SEQ ID NO:125, SEQ ID
NO:127, SEQ ID NO:129, SEQ ID NO:131, and SEQ ID NO:133.

. CA 02891714 2015-05-15

- 195 -E85. A nucleic acid encoding the antibody as set forth in any one of E77-E82, wherein the VH region is encoded by a nucleic acid comprising SEQ ID NO:110, and the VL region is encoded by a nucleic acid comprising SEQ ID NO:123.
E86. A vector comprising the nucleic acid as set forth in any one of E83-E85.
E87. A vector comprising a nucleic acid encoding the antibody as set forth in any one of E83-E86 E88. A cell comprising the nucleic acid as set forth in any one of E83-E85.
E89. A cell comprising the vector as set forth in any one of E86-E87.
E90. A cell expressing the antibody as set forth in any one of E77-E82.
E91. A cell comprising the nucleic acid as set forth in any one of E83-E85.
E92. A method of generating an antibody, comprising culturing the cell as set forth in any one of E88-E91 under conditions conducive to antibody expression, and allowing said cell to express said antibody.
E93. A pharmaceutical composition comprising the antibody as set forth in any one of E77-E82 and a pharmaceutically acceptable carrier.
E94. The antibody as set forth in any one of E1-71, and E77-E82.
The invention thus has been disclosed broadly and illustrated in reference to representative embodiments described above. Those skilled in the art will recognize that various modifications can be made to the present invention without departing from the spirit and scope thereof. All publications, patent applications, and issued patents, are herein incorporated by reference to the same extent as if each individual publication, patent application or issued patent were specifically and individually indicated to be incorporated by reference in its entirety. Definitions that are contained in text incorporated by reference are excluded to the extent that they contradict definitions in this disclosure.
It is appreciated that certain features of the invention, which are, for clarity, described in the context of separate embodiments, may also be provided in combination in a single embodiment. Conversely, various features of the invention which are, for brevity, described in the context of a single embodiment, may also be = CA 02891714 2015-05-15

- 196 -provided separately or in any suitable sub-combination.
It is specifically contemplated that any limitation discussed with respect to one embodiment of the invention may apply to any other embodiment of the invention.
Furthermore, any composition of the invention may be used in any method of the invention, and any method of the invention may be used to produce or to utilize any composition of the invention. In particular, any aspect of the invention described in the claims, alone or in combination with one or more additional claims and/or aspects of the description, is to be understood as being combinable with other aspects of the invention set out elsewhere in the claims and/or description and/or sequence listings and/or drawings.
In so far as specific examples found herein do not fall within the scope of an invention, said specific example may be explicitly disclaimed.
The use of the term "or" in the claims is used to mean "and/or" unless explicitly indicated to refer to alternatives only or the alternatives are mutually exclusive, although the disclosure supports a definition that refers to only alternatives and "and/or." As used herein the specification, "a" or "an" may mean one or more, unless clearly indicated otherwise. As used herein in the claim(s), when used in conjunction with the word "comprising", the words "a" or "an" may mean one or more than one.
As used herein "another" may mean at least a second or more. Unless otherwise defined herein, scientific and technical terms used in connection with the present invention shall have the meanings that are commonly understood by those of ordinary skill in the art. Further, unless otherwise required by context, singular terms shall include pluralities and plural terms shall include the singular. The words "comprises/comprising" and the words "having/including" when used herein with reference to the present invention are used to specify the presence of stated features, integers, steps or components but does not preclude the presence or addition of one or more other features, integers, steps, components or groups thereof.
Although the disclosed teachings have been described with reference to various applications, methods, and compositions, it will be appreciated that various

- 197 -changes and modifications can be made without departing from the teachings herein and the claimed invention below. The examples are provided to better illustrate the disclosed teachings and are not intended to limit the scope of the teachings presented herein. While the present teachings have been described in terms of these exemplary embodiments, numerous variations and modifications of these exemplary embodiments are possible without undue experimentation. All such variations and modifications are within the scope of the current teachings.
Where aspects or embodiments of the invention are described in terms of a Markush group or other grouping of alternatives, the present invention encompasses not only the entire group listed as a whole, but each member of the group individually and all possible subgroups of the main group, but also the main group absent one or more of the group members. The present invention also envisages the explicit exclusion of one or more of any of the group members in the claimed invention.
All references cited herein, including patents, patent applications, papers, text books, and the like, and the references cited therein, to the extent that they are not already, are hereby incorporated by reference in their entirety. In the event that one or more of the incorporated literature and similar materials differs from or contradicts this application, including but not limited to defined terms, term usage, described techniques, or the like, this application controls.
The description and examples detail certain specific embodiments of the invention and describes the best mode contemplated by the inventors. It will be appreciated, however, that no matter how detailed the foregoing may appear in text, the invention may be practiced in many ways and the invention should be construed in accordance with the appended claims and any equivalents thereof.

¨ 198-1st Component 2'd Component Antibody LC VL CL HC VH CH CH1 HNGE CH2 CH3 LC VL CL HC VH CH CH1 Abl 9 54 1 42 45 18 N/A N/A N/A N/A N/A
N/A N/A N/A N/A N/A
_ Ab2 9 _ 54 1 42 45 18 N/A
N/A N/A N/A N/A N/A , N/A N/A N/A N/A
-Ab1 Ab2 v1.0 24 _ 5533 42 45 46 _ Ab1 Ab2 v2.0 24 55 33 42 , 45 46 _ -Ab1 Ab2 v1.0 NEGATIVE 9 56 1 42 45 46 ¨
Abl Ab2 v1.0 T1 9 56 1 42 45 46 , 26 _ 61 35 42 45 47 -Abl Ab2 v1.0 T2 9 56 1 42 45 46 _ Ab1 Ab2 v1.0 T3_ 9 56 - 1 42 45 46 , _ _ Abl Ab2 v1.0 T4 9 56 1 42 45 46 _ . - -Ab1 Ab2 v1.0 T9 _ 9 56 1 42 45 46 _ 30 65 39 42 45 47 -Ab1 Ab2 v1.0 T12 9 56 1 42 45 46 -Abl Ab2 v1.0 T18 9 56 1 42 45 46 32- 67 41 42 45 47 r) -4=1 Ab1 Ab2 v1.0 T1* 24 , 55 33 42 45 46 , _ 26 61 35 42 45 47 o . -1..) Ab1 Ab2 v1.0 T2* 24 55 33 42 , 45 46 27 62 36 42 45 47 co -l0 Ab1 Ab2 v1.0 T3* 24 55 33 42 45 _ 46 , _ ...1 Ab1 Ab2 v1.0 T4* _ 24 55 33 42 45 46 _ _ io.
Ab1 Ab2 v1.0 T9* _ 24 55 33 42 , 45 46 . _ 1..) N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A o _ -1-, Ab3 9 54 1 42 45 18 , N/A N/A
N/A N/A N/A N/A N/A N/A N/A N/A
. i _ Ab3 C5-M1 _ 24 55 33 42 45 46 77 52 25 71 50 _ 59 34 42 45 47 0 in I
Ab3 C5-M1-NEGATIVE _ 9 56 1 42 45 46 76 _ 1-, Ab3 C5-M2 24_ 57 33 44 45 , 49 77 52 25 73 50 , 68 34 43 45 48 u, Ab3 C5-M2-NEGATIVE 9 58 1 44 45 49 76 52 9 N/A N/A
_ Deconvolute-3 (Ab3,C5) x86 x91 x89 42 45 46 x88 52 x87 x93 50 x92 x90 42 45 47 Deconvolute-4 (Ab3,C5) x24 x91 _ x89 42 45 46 x77 52 x25 x93 50 x92 x90 42 45 47 _ .
Ab3-S1xC5-T1 24 55 33 42 45 46 94 . -Ab3-51xC5-T2 24 55 33 42 45 46 95 Ab3-51xC5-T3 _ 24 5533 42 45 46 96 _ _ Ab3-51xC5-T4 _ 24 5533 42 45 46 97 - -Ab3-S1xC5-T9 _ 24 5533 42 45 _ 46 98 _ _ Ab3-S1xC5 24 55 33 42 45 46 76 Ab3xC5-S1rev 9 56 1 42 45 46 77 Table 26 SEQ IDs of antibodies and antibody domains used in examples. Each row indicates the seq IDs of sequences (see Table 1) used to construct the full sequence of the protein in column 1. (CH3). N/A: Not Applicable (bivalent, monospecific control antibody). Empty cells indicate unique sequence not disclosed in this application version.

Claims (30)

WHAT IS CLAIMED IS:

1. A heterodimeric protein, comprising (i) a first C H1 domain (C H1) and a first C L domain (C L), the first C H1 and the first C L interacting together at a first C H C L interface to form a first C H C L domain (C H C L);
(ii) a second C H1 domain (C H1) and a second C L domain (C L), the second C H1 and the second C L interacting together at a second C H C L interface to form a second C H C L domain (C H C L);
wherein the first C H1 is engineered to differ from the second C H1 by at least one C H1 mutant residue in the first C H1; and the first C L is engineered to differ from the second C L by at least one C L mutant residue in the first C L;
such that the C H1 mutant residue and the C L mutant residue of the first C H
C L
interact with each other in preference to the corresponding residue positions on the second C H C L, the interacting mutant residues of the first C H1 and first C L thereby forming a first complementary residue set, and wherein the first C H1 is attached to a first variable heavy domain (V H), and the first C L is attached to a first variable light domain (V L), and the second C H1 is attached to a second V H, and the second C L is attached to a second V L, such that when combined, the first V H, first V L, first C H and first C L
together form a first Fab, and when combined, the second V H, second V L, second C H1, and second C L form a second Fab, and wherein preferential formation of the first Fab and the second Fab does not rely on complementary pairing of the variable domains.

2. The heterodimeric protein as claimed in claim 1, wherein the solvent accessible surface area of the first complementary residue set is less than 225.ANG.2 as measured using a 2.5.ANG. probe.

3. The heterodimeric protein as claimed in claim 1 or 2, wherein the second C

is engineered to differ from the first C H1 by at least one C H1 mutant residue in the second C H1; and the second C L is engineered to differ from the first C L
by at least one C L mutant residue in the second C L; such that the C H1 mutant residue and the C L mutant residue of the second C H C L preferentially interact with each other over the corresponding residue positions on the first C H C L, the interacting mutant residues of the second C H1 and second C L thereby forming a second complementary residue set.

4. The heterodimeric protein as claimed in claim 3, wherein the solvent accessible surface area of the second complementary residue set is less than 225.ANG.2 as measured using a 2.5.ANG. probe.

5. The heterodimeric protein as claimed in any one of claims 2-4, wherein the mutant residues of the first complementary residue set are different to the mutant residues of the second complementary residue set.

6. The heterodimeric protein as claimed in any one of claims 1-5, wherein formation of the first C H C L and second C H C L preferentially occurs over formation of a C H C L comprised of either the first C H1 and second C L, or second C H1 and first C L.

7. The heterodimeric protein as claimed in claim 6, wherein formation of the first C H C L and second C H C L preferentially occurs over formation of a C H C L
comprised of either the first C H1 and second C L, or second C H1 and first C
L by at least about 4-fold.

8. The heterodimeric protein as claimed in claim 6, wherein formation of the first C H C L and second C H C L preferentially occurs over formation of a C H C L
comprised of either the first C H1 and second C L, or second C H1 and first C
L by at least about 10-fold.

9. The heterodimeric protein as claimed in claim 6, wherein formation of the first C H C L and second C H C L preferentially occurs over formation of a C H C L
comprised of either the first C H1 and second C L, or second C H1 and first C
L by at least about 20-fold.

10.The heterodimeric protein as claimed in any one of claims 1-9, wherein at least one of the C L domains is a kappa domain.

11.The heterodimeric protein as claimed in any one of claims 1-10, wherein the complementary residue sets comprise a positively or negatively charged residue in one domain, and either a polar residue, or an oppositely charged residue in the other domain.

12.The heterodimeric protein, as claimed in any one of claims 1-11, wherein the locations of the complementary residue sets are selected from the group consisting of:
(i) C H1-124 and C L-176;
(ii) C H1-188 and C L-178;
(iii) C H1-143 and C L-178;
(iv) C H1-143 and C L-131;
(v) C H1-221 and C L-123;
(vi) C H1-145 and C L-131;
(vii) C H1-179 and C L-131;
(viii) C H1-186 and C L-131; and (ix) C H1-188 and C L-133.

13.The heterodimeric protein, as claimed in claim 12, wherein either the mutation at the C H1 position is selected from the group consisting of W, H, K, R, S
and T, and the mutation at the C L position is selected from the group consisting of S, M, D and E, or the mutation at the C H1 position is selected from the group consisting of E, and D, and the mutation at the C L position is selected from the group consisting of H, K, and R.

14.The heterodimeric protein, as claimed in claim 13, wherein the complementary residue sets further comprise one or more mutations selected from the group consisting of: C H1-143D, C H1-145S, C H1-186A, C H1-186E,C H1-188G, C H1-143S, C H1-190S, C H1-190I, C L-133S, C L-135I, C L-176G, C L-176M, and C L-178G, C
L-178S.

15.The heterodimeric protein, as claimed in any one of claims 1-14, wherein the first and second complementary residue sets are selected from two of the following groups:
(i) C H1-124K, C L-176D, C H1-190S, C L-133S;
(ii) C H1-124K, C L-176D, C L-133S;
(iii) C H1-124E, C L-176K;

(iv) C H1-124E, C L-176K, C H1-188G;
(v) C H1-188E, C L-178K, C H1-143E;
(vi) C H1-188K, C L-178D, C H1-143D;
(vii) C H1-143K, C L-178D;
(viii) C H1-143D, C L-178R;
(ix) C H1-143K, C L-178D;
(x) C H1-143D, C L-178K;
(xi) C H1-143D, C L-178K, C H-176M;
(xii) C H1-143E, C L-131R;
(xiii) C H1-143R, C L-131E;
(xiv) C H1-143R, C L-131E, C H1-186A;
(xv) C H1-221D, C L-123K;
(xvi) C H1-221D, C L-123K, C H1-190I, C L-135I;
(xvii) C H1-145E, C L-131H;
(xviii) C H1-143H, C H1-179D, C H1-186E, C L-131H;
(xix) C H1-145E, C L-131H;
(xx) C H1-186E, C L-131H, C H1-145S;
(xxi) C H1-143S, C L-131D, C H1-188W, C L-133S, C L-178S;
(xxii) C H1-143S, C H1-188W, C L-133M, C L-176G, C L-178G;
(xxiii) C H1-143H, C H1-179D, C H1-186E, C L-131H, C H-190I, C L-135I,;
(xxiv) C H-186E, C L-131H, C H-145S;
(xxv) C H1-143S, C L-131D, C H1-188W, C L-133S, C L-176C;
(xxvi) C H1-143S, C H1-188W, C L-133M, C L-178G, C L-176G;
(xxvii) C H1-143S, C H1-188W, C L-131D.

16.The heterodimeric protein, as claimed in any one of claims 1-15, comprising an engineered disulfide bond between the first C H1 and the first C L, and or the second C H1 and the second C L.

17.The heterodimeric protein, as claimed in claim 16, wherein the engineered disulfide bond is located at one or more of the following positions:
(i) C H1-122 and C L-123;

(ii) C H1-139 and C L-116; and (iii) C H1-174 and C L-176.

18.The heterodimeric protein, as claimed in claim 16 or 17, wherein a wild type disulfide bond has been removed, by mutating one or both of C H1-C230 and C L-214 to any residue except C, on the first C H C L and/or second C H C L, and wherein the first and/or second C H1-C230 and first, and/or second C L-C214 are mutated to S.

19.The heterodimeric protein, as claimed in any one of claims 1-18, wherein the first C H C L comprises residues from one of the following groups:
(i) C H1-124K, C L-176D, C H1-190S, C L-133S;
(ii) C H1-124K, C L-176D, C L-133S;
(iii) C H1-124E, C L-176K, C L-133S;
(iv) C H1-124E, C L-176K, C H1-188G, C L-133S;
(v) C H1-188E, C L-178K, C H1-143E;
(vi) C H1-188K, C L-178D, C H1-143D;
(vii) C H1-143K, C L-178D;
(viii) C H1-143D, C L-178R;
(ix) C H1-143K, C L-178D;
(x) C H1-143D, C L-178K;
(xi) C H1-143D, C L-178K, C L-176M;
(xii) C H1-143E, C L-131R;
(xiii) C H1-143R, C L-131E;
(xiv) C H1-143R, C L-131E, C H1-186A, (xv) C H1-221D, C L-123K;
(xvi) C H1-221D, C L-123K, C H1-190I, C L-135I, C H1-174C, C H1-230S, C L-176C, C L-214S;
(xvii) C H1-145E, C L-131H;
(xviii) C H1-143H, C H1-179D, C H1-186E, C L-131 H;
(xix) C H1-122C, C H1-145E, C H1-230S, C L-123C, C L-131H, C L-214S;
(xx) C H1-186E, C L-131H, C H1-145S;

(xxi) C H1-143S, C L-131D, C H1-188W, C L-133S, C L-178S;
(xxii) C H1-143S, C H1-188W, C L-133M, C L-176G, C L-178G;
(xxiii) C H1-143H, C H1-179D, C H1-186E, C L-131H, C H-190I, C L-135I, C H1-174C , C H1-230S, C L-176C , C L-214S;
(xxiv) C H-186E, C L-131H, C H-145S, C H1-139C , C H1-230S, C L-116C , C L-214S;
(xxv) C H1-143S, C L-131D, C H1-188W, C L-133S, C L-178S, C H1-174C , C H1-230S, C L-176C , C L-214S;
(xxvi) C H1-143S, C H1-188W, C H1-122C , C H1-230S, C L-133M, C L-178G, C L-176G, C L-123C , C L-214S;
(xxvii) C H1-143S, C H1-188W, C H1-122C , C H1-139C , C H1-174C , C H1-230S, C
L-133S, C L-178S, C L-131D, C L-116C , C L-123C , C L-176C , C L-214S.

20.The heterodimeric protein, as claimed in claim 19, wherein the second C HC
L
comprises residues from one of groups (i)-(xxvii), provided the first and second C H C L do not both comprises residues from the same group.

21.The heterodimeric protein as claimed in any one of claims 1-20, wherein the first C H1 is connected to a first C H2 domain (C H2), which is connected to a first C H3 domain (C H3), and the second C H1 is connected to second C H2, which is connected to a second C H3, such that the first C H3 and second C H3 comprise a first C H3 mutant residue and second C H3 mutant residue respectively, the first C H3 mutant residue and second C H3 mutant residues being engineered to differ from each other, and preferentially interact with each other and thereby form C H3 heterodimers over the formation of C H3 homodimers.

22.A bispecific antibody comprising a heterodimeric protein as claimed any one of claims 1-21, wherein the first C H C L comprises C H1-124K, C L-176D, C H1-190S, and C L-133S, and the second C H C L comprises C H1-124E, C L-176K, C H1-188G, and C L-133S.

23.A nucleic acid encoding the heterodimeric protein as claimed in any one of claims 1-21.

24.A vector comprising the nucleic acid as claimed in claim 23.

25.A cell comprising the nucleic acid as claimed in claim 24, or comprising the vector as claimed in claim 22.

26.A method of making the heterodimeric protein, as claimed in any one of claims 1-21 comprising:
(i) cotransfecting a cell line with one or more vectors to express the first C H1, the first C L of the first C H C L; and the second C H1, and the second C L of the second C H C L;
(ii) culturing the cell line under conditions to express the one or more vectors and that allow the first C H C L and second C H C L to assemble;
and (iii) purifying the heterodimeric protein from the cell culture.

27.The method of claim 26, wherein the cell line is cotransfected with vectors that express the first C H1, first C L, second C H1, and second C L in a 1:1:1:1 ratio.

28. An isolated antibody that specifically binds human TrkB, wherein the antibody comprises a VH region comprising the amino acid of SEQ ID NO:51 and a VH
region comprising the amino acid sequence of SEQ ID NO:53.

29. The antibody of claim 28, wherein the heavy chain comprises the amino acid sequence of SEQ ID NO:75 and the light chain comprises the amino acid sequence of SEQ ID NO:78.

30.A pharmaceutical composition comprising the antibody of claim 28 and a pharmaceutically acceptable carrier.
3I.An isolated nucleic acid encoding the antibody of claim 28.